Open Agentic AI And True Personalisation

Sixteen years on from my own initial posts on the subject of a personal assistant that can privately find and buy stuff for you on the web, and we have 'open agentic AI'. But are you really any closer to the automated selection and purchase of your own personalised products without needlessly surrendering your privacy or otherwise becoming the victim? Should this latest iteration of open generative AI be autonomously making and executing decisions on your behalf? 

What is Agentic AI?

An 'agentic' AI is an evolution of generative AI beyond a chatbot. It receives your data and relies on pattern matching to generate, select and execute one of a number of potential pre-programmed actions without human guidance, then 'learns' from the result (as NVIDIA, the leading AI chip maker, explains). 

A 'virtual assistant' that can find, buy and play music, for example, is a form of agentic AI (since it uses AI in its processes), but the ambition involves a wider range of tasks and more process automation and autonomy (if not end-to-end). 

You'll see a sleight-of-hand in the marketing language (like NVIDA's) as developers start projecting 'perception', 'understanding' and 'reasoning' on their agentic AIs, but computers don't actually do any of those human things. 

It's certainly a compelling idea to apply this to automating various highly complex, tedious consumer 'workflows' that have lots of different parameters - like buying a car, perhaps (or booking a bloody rail ticket in the UK!). 

Wearing my legal hat, I also see myriad interesting challenges (which I'd be delighted to discuss, of course!), some of which are mentioned here, but not all...

Some challenges

The main problem with using an 'agentic AI' in a consumer context is the involvement of a large language model and generative AI where there is a significant (e.g. economic, medical and/or legal) consequence for the user (as opposed to a chatbot or information-only scenario (though that can also be problematic). Currently, the household or device based virtual assistants are carrying out fairly mundane tasks, and you could probably get a refund if it bought you the wrong song, for example, if that really bothered you. Buying the wrong car would likely be a different matter.

There may also be confusion about the concept of 'agency' here. The word 'agentic' is used to mean that the AI has 'agency' in the sense it can operate without human guidance. That AI is not necessarily anyone's legal 'agent' (more below) and is trained on generic training data (subject to privacy, copyright consents/licensing), which these days is itself synthetic - generated by an AI. So, agentic AIs are not hosted exclusively by or on behalf of the specific consumer and do not specifically cater to a single end-customer's personalised needs in terms of the data it holds/processes and how it deals with suppliers. It does not 'know' you or 'understand' anyone, let alone you.  

Of course, that is consistent with how consumer markets work: products have generally been developed to suit the supplier's requirements in terms of profitability and so on, rather than any individual customer's needs. Having assembled what the supplier believes to be a profitable product by reference to an ideal customer profile in a given context, the supplier's systems and marketing/advertising arrangements seek out customers for the product who are 'scored' on the extent to which they fit that 'profile' and context. This also preserves 'information asymmetry' in favour of the supplier, who knows far more about its product and customers than you know about the supplier or the product. In an insurance context, for example, that will mean an ideal customer will pay a high premium but find it unnecessary, too hard or impossible to make a claim on the policy. For a loan, the lender will be looking for a higher risk customer who will end up paying more in additional interest and default fees than lower risk customers. But all this is only probabilistic, since human physiology may be 'normally distributed' but human behaviour is not.

So using an agentic AI in this context would not improve your position or relationship with your suppliers, particularly if the supplier is the owner/operator of the agentic AI. The fact that Open AI has offered its 'Operator' agentic AI to its pro-customers (who already pay a subscription of $200 a month!) begs the question whether Open AI really intends rocking this boat, or whether it's really a platform for suppliers like Facebook or Google search in the online advertising world. 

It's also an open question - and a matter for contract or regulation - as to whether the AI is anyone's legal 'agent' (which it could be if the AI were deployed by an actual agent or fiduciary of the customer, such as a consumer credit broker). 

Generative AI also has a set of inherent risks. Not only do they fail to 'understand' data, but to a greater or lesser degree they are also inaccurate, biased and randomly hallucinate rubbish (not to mention the enormous costs in energy/water, capital and computing; the opportunity cost of diverting such resources from other service/infrastructure requirements; and other the 'externalities' or socioeconomic consequences that are being ignored and not factored into soaring Big Tech stock prices - a bubble likely to burst soon). It may also not be possible to explain how the AI arrives at its conclusions (or, in the case of an agentic AI, why it selected a particular product, or executed a specific task, rather than another). Simply overlaying a right to human intervention by either customer or supplier would not guarantee a better outcome on theses issues (due to lack of explainability, in particular). A human should be able to explain why and how the AI's decision was reached and be able to re-take the decision. And, unfortunately, we are seeing less and less improvement in each of these inherent risk areas with each version of generative AIs.

All this means that agentic AI should not be used to fully automate decisions or choices that have any significant impact on an individual consumer (such as buying a car or obtaining a loan or a pension product).  

An Alternative... Your Own Personal Agent

What feels like a century ago, in 2009, I wondered whether the 'semantic web' would spell the end of price comparison websites. I was tired of seeing their expensive TV ads - paid for out of the intermediary's huge share of the gross price of the product. I thought: "If suppliers would only publish their product data in semantic format, a 'widget' on my own computer could scan their datafeeds and identify the product that's right for me, based on my personal profile and other parameters I specify". 

By 2013, I was calling that 'widget' an Open Data Spider and attempted to explain it further in an article for SCL on the wider tech themes of Midata, Open Data and Big Data tools (and elsewhere with the concept of 'monetising you'). I thought then - and still think now - that: 

"a combination of Midata, Open Data and Big Data tools seems likely to liberate us from the tyranny of the 'customer profile' and reputational 'scores', and allow us instead to establish direct connections with trusted products and suppliers based on much deeper knowledge of our own circumstances."

Personalised assistants are evolving to some degree, in the form of 'personal [online] data stores' (like MyDex or Solid); as well as 'digital wallets' or payment apps that sit on smartphones and other digital devices and can be used to store transaction data, tickets, boarding passes and other evidence of actual purchases. The former are being integrated in specific scenarios like recruitment and healthcare; while the latter tend to be usable only within checkout processes. None seems to be playing a more extensive role in pre-evaluating your personal requirements, then seeking, selecting and purchasing a suitable product for you from a range of potential suppliers (as opposed to a product that a supplier has created for its version of an 'ideal' customer that you seem to fit to some degree). 

Whether the providers of existing personal data stores and digital wallets will be prepared to extend their functionality to include more process automation for consumers may also depend on the willingness of suppliers to surrender some of their information advantage and adapt their systems (or AIs) to respond to and adapt products according to actual consumer requests/demand.

Equally, the digital 'gatekeepers' such as search providers and social media platforms will want to protect their own advertising revenue and other fees currently paid by suppliers who rely on them for targeting 'ideal' customers. Whether they can 'switch sides' to act for consumers and preserve/grow this revenue flow remains to be seen.

Overall, if I were a betting man, I'd wager that open agentic AI won't really change the fundamental relationship between suppliers, intermediaries and consumers, and that consumers will remain the targets (victims) for whatever suppliers and intermediaries dream up for them next...

I'd love to be corrected!

You Pay For Social Media, AI and Crypto Via Your Utility Bills

Households consume the most electricity in the UK. That's why the huge surge in global energy prices following the expansion of Russia's invasion of Ukraine in February 2022 prompted the government to invoke price caps and handouts to protect consumers (and businesses) from bankruptcy, along with private energy providers who failed to manage their market exposures. Yet few have noticed that the world's computing data centres, including those hosting artificial intelligence platforms, already consume enough energy to power entire countries and compete with humans for vast amounts of fresh water. Crypto-currencies also require huge amounts of energy to 'mine'. So, not only are you paying for many 'free' online services with your personal data, you're also paying through your energy and water bills. And based on the advertising and other revenues from your participation, Big Tech can afford to outspend you. To illustrate the challenge, the UK government just announced a massive new Microsoft AI facility in London, even as Thames Water circles the drain and lack of capacity in the UK's national grid is delaying the construction of new homes, renewable energy projects and electric vehicle charging points. Given these costs and shortages, should we be speculating in bitcoin and using generative AI (either for fun or to do things we could do for ourselves)?

How much power does the latest consumer technology use?

While consumer electronics only account for 6% of household usage, that doesn't account for the centralised data processing among digital media and gaming platforms, for example, when you participate online. As a result, households are responsible for 35% of electricity usage, services 29% and industry 30%. You might argue that much of this data centre capacity is used by businesses, but many of them do so ultimately to serve consumers - from online search, shopping and social media services to powering giant credit card networks. 

Artificial intelligence, however, operates at a whole new level above the more traditional digital media. A Netflix fan would have to have watched 1,625,000 hours of content to use the same amount of power it took to train OpenAI's ChatGPT 3.0 during 2022, according to a Dutch researcher. Generating a single image from text on other AI platforms costs the same amount in energy as charging your smartphone.

The same Dutch researcher has estimated that the AI sector alone will use as much power as the Netherlands by 2027, while the International Energy Agency predicts that the world's data centres (including AI and other digital media) will consume the double the amount of electricity in 2026 that they consumed in 2022 - about as much as Japan (the 5th largest electricity consumer in the world, behind China, the US, India and Russia).  

Bitcoin mining - an activity whose sole purpose is to feed the world's first and largest distributed Ponzi scheme - absorbed nearly 1% of the world's electricity in 2023 - enough to power Greece or Australia. That's up to 5 times the cost of legacy payment systems that process vastly more transactions (though they also use enough to electricity to power Portugal or Bangladesh).

How much water does the latest consumer technology use?

Data centres also consume vast amounts of water (not counting what they recycle) to cool the computers and humidify the internal air. But even the process of generating the electricity they use also consumes water. 

In 2021, for example, Google's data centers consumed approximately 4.3 billion gallons of water (16.3 billion litres), an average of 450,000 gallons (1.7m litres) of water per data centre each day. Microsoft reckoned that it consumed 1.7 billion gallons (nearly 6.5 billion litres) in 2022.

Gridlock

The surge in energy and water usage by future-gazing tech providers comes at a time when Britain's infrastructure is already failing to support the construction of new and more energy efficient homes, renewable energy sources and the switch away from diesel and petrol vehicles.

“Nationally, we’ve got an absolute ­crisis in all infrastructure.” Plans by Michael Gove, the housing secretary, to build 150,000 homes in Cambridge to create a British Silicon Valley were already being hampered by lack of water... “And where’s the power coming from? Something fundamental has to change...”

"...90 new homes in the Littlemore district had been meant to have heat pumps. “The National Grid basically said ‘we won’t have enough power to connect them’ so half the houses are going to have to have gas boilers instead – it’s so frustrating. 

Great Britain’s power stations together generate 75 gigawatts of electricity, and the mainland is expected to need about twice as much by 2050 as people switch to ­electric vehicles and heat pumps.” The Guardian

Dissatisfaction with Britain's electric vehicle charging network is running at about 70% of EV drivers, citing a lack of public charging stations and unreliability. The government is targeting 300,000 charging stations by 2030, with only 53,677 available at the start of 2024 (an increase of 45% in 12 months) and the majority to be provided by private investors.

Meanwhile, Britain's water problems flow partly from the risk of drought and party from its combined sewage system which takes rainwater through the same pipes as the grey water from sinks and baths, as well as the raw sewage from toilets. Any excess of rainwater simply overloads the sewerage system of pipes that normally takes sewage to local treatment works, and the overflow goes directly into the waterways... 

Crisis? What Crisis?

Who's to blame for Britain's sagging infrastructure involves lots of finger-pointing and misinformation. 

When challenged over delays to connect new systems to the electricity grid, the National Grid's system operator complains that the queue of projects waiting to connect would add 800 gigawatts of electricity - "more than more than four times as much as the country would ever need." There are even delays in the time it takes to get an estimate of when a project will be connected, as well as 'zombie projects' that were approved but have been abandoned due to connection waiting times of 5 to 15 years.

Yet this hides the fact that more renewable projects/systems will be necessary to reduce Britain's reliance on fossil fuels, since energy systems that generate electricity from solar and wind don't all contribute to the grid at the same time, unlike a gas-fired or coal-fired power station where the energy source to create the electricity is under human control. 

As for water - well, none of England’s rivers is classified as being in good ecological health and Britain is already failing to produce enough fresh water to meet its needs year round. The country's 'combined sewage system' should be replaced by separate systems for rainwater and sewage, yet modernisation efforts have merely doubled-down on the combined system.

UK Government Distracted by Culture Wars

Britain's energy sector is self-evidently poorly prepared for the future. Here is a good description of the alphabet soup of bodies involved and the problem of every additional significant energy system creating the need for some change in part of the network. Here's a good overview of the challenges facing sewerage reform and here is a discussion of drought risk.

There is undoubtedly a need for reforms and there are plenty that have been announced with targets of, say, 2035 and 2050, but where are the plans that had a target of, say, 2023? And if we had them, why weren't they being updated?

It's hardly surprising that a country with 5 prime ministers in 8 years and as many Cabinet reshuffles has failed to find the time or dedication required to overhaul the energy sector and water industry. Too much control over the maintenance and renewal of Britain's creaking infrastructure has been left to private interests. Had the income from customer bills gone into public coffers instead of draining into investors' pockets, it might have been a different story - or at least the money might have been used to bolster the many other public services that are in such a dire state. 

Choices, Choices

All this brings us back to scarcity and the need to make careful choices over how we develop, protect and deploy our energy and water resources. This is largely a question of politics and intervention by a responsible government to balance out the many competing interests. Areas in which Brexit Britain has been - and continues to be - very poorly served. 

It must be doubted that a new government will be able to make much progress after 15 years of under-investment and poor decision-making by its predecessors.

In these circumstances, it seems unwise to devote enormous amounts of power and water to mine bitcoin for speculative purposes or to support generative AI systems that are either used merely for entertainment or to render people jobless (if the hype is to be believed). 

Certainly cash-strapped consumers should think about their utility bills and water shortages before speculating in bitcoin, playing online games or using open AI systems for entertainment or to do things that they could do for themselves. 

 

There's Nothing Intelligent About The Government's Approach To AI Either

Surprise! The UK government's under-funded, shambolic approach to public services also extends to the public sector's use of artificial intelligence. Ministers are no doubt piling the pressure on officials with demands for 'announcements' and other soundbites. But amid concerns that even major online platforms are failing to adequately mitigate the risks - not to mention this government's record for explosively bad news - you'd have thought they'd tread more carefully.

Despite 60 of the 87 public bodies either using or planning to use AI, the National Audit Office reports a lack of governance, accountability, funding, implementation plans and performance measures. 

There are also "difficulties attracting and retaining staff with AI skills, and lack of clarity around legal liability... concerns about risks of unreliable or inaccurate outputs from AI, for example due to bias and discrimination, and risks to privacy, data protection, [and] cyber security." 

The full report is here.

Amid concerns that the major online platforms are also failing to adequately mitigate the risks of generative AI (among other things), you'd have thought that government would be more concerned to approach the use of AI technologies responsibly.

But, no...

For what it's worth, here's my post on AI risk management (recently re-published by SCL).

AI is a Set of Technologies, Not The End Of Work

We've heard a lot for a long time about artificial intelligence replacing our jobs and, ultimately, the human race. We're told we'll need to retrain to do things that AI computers cannot. But beware the hype. After all, AI is just a set of technologies and we've coped with the introduction of new technology before. Rather than having to retrain, it's more likely you'll be using AI without even realising it. And there are cases where robotics are needed because humans are reluctant or unavailable to do certain tasks... The real concern is that the hype distracts us from more immediate and greater threats posed by AI and how to manage and regulate the risks appropriately.

Much of the hype surrounding AI confuses its development with its actual or potential uses, whether in business or in the course of other activities, like writing a wedding speech. As with any technology, there's obviously a business in developing AI, selling it and deploying it. But how useful it is depends on who uses it and how.

This confusion is perhaps partly driven by the fact that some businesses are developing and operating 'open' AI systems-as-a-service focused on particular use-cases or scenarios (chatbots, research, text-to-image and so on), so you conduct your activity on their platform rather than your own device. The hype surrounding these platforms is intended to attract investment and users, but it seems unlikely that they will become the Alphabet (Google), Microsoft or Meta (Facebook) of tomorrow, especially as those tech giants are funding AI development themselves, to cement their own market dominance. 

Yet, while the tech giants might dominate the markets for their technology (and some markets where they act as more than just a tech business, like advertising), you'll see that they aren't dominating every business sector or industry in which their technology is used. 

It's therefore the people and businesses who successfully deploy and use AI who will benefit from those technologies, not the developers. This is no different to the use of telephones, laptop computers or email (or a distributed ledger supporting a cryptocurrency). 

Nobody who went from using a typewriter or the analog version of a telephone, laptop, email or work intranet would say that they're redundant or even work less as a result of using the new/replacement technology. If anything, those tools have enabled changes in work patterns that have meant that humans work faster, longer and, ultimately, harder.

And there was more 'retraining' involved in introducing PCs, email, spreadsheets and video conferencing than AI, which may be so embedded in existing processes that you don't even realise you're using it, whether in terms of product recommendations, chatbots and virtual assistants, predictive text and search features, or tagging your friends in the social media. 

There is plenty of speculation that truck drivers will be replaced by robots. Maybe truck technology has evolved to mean fewer drivers per tonne of truck, but there has been a steady increase in the number of trucks (and therefore demand for drivers) in the UK, for example, and driving a giant HGV takes more skill than smaller vehicles. Yet, ironically, there is a persistent shortage of drivers, so that transport firms are effectively being forced to invest in autonomous vehicles, just as farmers are turning to robotics due to a shortage of humans willing to pick fruit and vegetables). There are also many risky tasks that are better done remotely by machines, such as working in radioactive or other dangerous environments. AI may still be a threat to those still willing to do those tasks, yet they could also benefit from the demand for experienced humans to help in the wider development, deployment and use of the robots. This is no different to previous waves of technological innovation.

Yet all humans have a genuine concern if their personal information is being included in an AI's training data or is otherwise being harvested and used without your consent. That's where humans need to focus urgently, as well as in the creative industries where copyright violation by AIs is also rife. We also need to be on guard against hallucinating AIs, disinformation, deepfakes and misinformation - particularly in an election year.

More on that soon...

Defending Humanity Against The Techno-Optimists

I've been involved in tech since the mid-90s, have experienced the rise and burst of many 'bubbles', and have been writing about SiliCon Valley's war on the human race since 2014. But the latest battles involving crypto and AI are proving to be especially dangerous. A cult of 'techno-optimism' has arisen, with a 'manifesto' asserting the dominance of their own self-interest, backed by a well-funded 'political action committee' making targeted political donations. Laws and lawsuits are pending, but humanity has to play a lot harder on defence... To chart a safe route, we must prioritize the public interest, and align technology with widely shared human values rather than the self-interest of a few tech enthusiasts, no matter how wealthy they are.

As Michael Lewis illustrated in The New New Thing, SiliCon Valley has always had its share of people eager to get rich flogging a 'minimum viable product' that leaves awkward 'externalities' for others to deal with. Twenty five years on, we are still wrestling with disinformation and other harmful content that flows from social media platforms, for example, never mind the 'dark web'.

Regardless of the potential downsides, the 'Techno-optimist manifesto' seeks to elevate and enshrine the get-rich-quick-at-others'-expense approach in a set of beliefs or 'creed' with technology as a 'god':

"Technology is the glory of human ambition and achievement, the spearhead of progress, and the realization of our potential." a16z

The techno-optimist creed commands followers to view the world only in terms of individual self-interest, to a point verging on malignant narcissism:

"We believe markets do not require people to be perfect, or even well intentioned – which is good, because, have you met people? Adam Smith: “It is not from the benevolence of the butcher, the brewer, or the baker that we expect our dinner, but from their regard to their own self-interest. We address ourselves not to their humanity but to their self-love, and never talk to them of our own necessities, but of their advantages.” a16z

In other words, techno-optimists aren't interested in humanity, good intentions or benevolence. They are only self-interested and believe that you and everyone else is, too. It's you against them, and them against you. In this way, the techno-optimists absolve themselves of any responsibility to care about other humans, because other humans are merely self-interested and technology is the pinnacle of everyone's self-interest. 

The cult only needs to focus on building new tech. 

The only remaining question relating to other humans is whether your self-interest is aligned with the techno-optimist's chosen technology. If not, you lose - as we'll see when it comes to their use of your cryptoassets or your copyright work or personal data where it is gathered among the training data they need to develop AI systems...

You might well ask if there are any constraints at all on the techno-optimists' ambition, and I would suggest only money, tech resources and the competing demands of other techno-optimists.

They claim not to be against regulation, so long as it doesn't throttle their unrestrained ambition or 'kill' their pet technology. To safeguard their self-interest, the techno-optimists are actively funding politicians who are aligned with their self-interest and support their technology, and attacking those who are not... with a dose of nationalism for good measure:

“If a candidate supports an optimistic technology-enabled future, we are for them. If they want to choke off important technologies, we are against them,” wrote Ben Horowitz, one of [a16z's] founders, in a Dec. 14 post, adding: “Every penny we donate will go to support like-minded candidates and oppose candidates who aim to kill America’s advanced technological future.” Cointelegraph

"Fairshake, a political action committee [PAC] supported by Coinbase and a16z, has a $73 million war chest to oppose anti-crypto candidates and support those in favor of digital assets... Fairshake describes itself as supporting candidates “committed to securing the United States as the home to innovators building the next generation of the internet.” Cointelegraph

Nationalistic claims are typical of such libertarian causes (Trump's "Make America Great Again") and invite unfortunate comparisons with European politics of the 1930s, as George Orwell pointed out in his Notes on Nationalism in 1945:

Nationalism is not to be confused with patriotism... two different and even opposing ideas are involved. By ‘patriotism’ I mean devotion to a particular place and a particular way of life, which one believes to be the best in the world but has no wish to force on other people. Patriotism is of its nature defensive, both militarily and culturally. Nationalism, on the other hand, is inseparable from the desire for power... 

A nationalist is one who thinks solely, or mainly, in terms of competitive prestige. He may be a positive or a negative nationalist — that is, he may use his mental energy either in boosting or in denigrating — but at any rate his thoughts always turn on victories, defeats, triumphs and humiliations. He sees history, especially contemporary history, as the endless rise and decline of great power units, and every event that happens seems to him a demonstration that his own side is on the upgrade and some hated rival is on the downgrade. 

But finally, it is important not to confuse nationalism with mere worship of success. The nationalist does not go on the principle of simply ganging up with the strongest side. On the contrary, having picked his side, he persuades himself that it is the strongest, and is able to stick to his belief even when the facts are overwhelmingly against him. Nationalism is power-hunger tempered by self-deception. Every nationalist is capable of the most flagrant dishonesty, but he is also — since he is conscious of serving something bigger than himself — unshakeably certain of being in the right..."

Yet in 2014, Google's CEO at the time, Eric Schmidt, 'warned' us that humans can only avoid the much vaunted Singularity - where computers out-compete humans to the point of extinction - by finding things that 'only humans can do and are really good at'. Ironically, by dedicating themselves utterly to the god of technology, the techno-optimist is actually asserting the 'self-interest' of machines! 

Of course, technology is not inherently good or bad. That depends on their human creators, deployers and users. There's a long list of problems in the techno-optimist manifesto which they claim technology itself has 'solved' but self-evidently has not, either because the technology was useless without human involvement or the problems persist.

And what of their latest creatures: crypto and AI?

While 'blockchain' or distributed ledger technology does have some decent use-cases, the one that gets the techno-optimists most excited is using crypto-tokens as either a crypto-currency or some other form of tradeable crypto-asset. They insist that the technology is so distinct that it must not be subject to existing securities laws. Yet they use the terminology of existing regulated markets to describe roles in the crypto markets that are really only corruptions of their 'real world' counterparts. Markets for cryptocurrencies and cryptoassets are riddled with examples of fraud and market manipulation that were long ago prohibited in the regulated markets. A supposedly distributed means of exchange without human intervention is actually heavily facilitated by human-directed intermediaries, some of which claim to operate like their real world equivalents that safeguard their customers' funds, while actually doing the opposite. The shining example of all these problems, and the numerous conflicts with the participating techno-optimists' self-interest, is the FTX scandal. And there are many others.

As for AI, again there are decent systems and use-cases, but the development of some AI systems relies on huge sets of 'training data' that would be prohibitively expensive to come by, were they not simply 'scraped' from the internet, regardless of copyright or privacy concerns: the technological equivalent of toxic waste. The creators of several of these 'open' AI systems defend their activity on techno-optimist grounds. Midjourney founder David Holz has admitted that his company did not receive consent for the hundreds of millions of images used to train its AI image generator, outraging photographers and artists; and OpenAI blithely explained in its submission to a UK House of Lords committee:

“Because copyright today covers virtually every sort of human expression – including blogposts, photographs, forum posts, scraps of software code, and government documents – it would be impossible to train today’s leading AI models without using copyrighted materials.”

So, there we were in 2014 being warned to be creative, but it turns out that the techno-optimists believe that your self-interest and the rights that protect your work can simply be overridden by their 'divine' self-interest. 

Needless to say, many humans are not taking this lying down (even if some of their governments and institutions are).

In January 2023, illustrators sued Midjourney Inc, DeviantArt Inc (DreamUp), and Stability A.I. Ltd (Stable Diffusion), claiming these text-to-image AI systems are “21st-century collage tools that violate the rights of millions of artists.”  A spreadsheet submitted as evidence allegedly lists thousands of artists whose images the startup's AI picture generator "can successfully mimic or imitate." 

The New York Times has sued OpenAI and Microsoft for copying and using millions of its copyright works seeking to free-ride on its investment in its journalism by using it to build 'substitutive' products without permission or payment.  

Getty Images has also filed a claim that Stability AI ‘unlawfully’ scraped millions of images from its website. 

Numerous other lawsuits are pending; and legislative measures have either been passed (as in the EU and China) or regulators have been taking action under existing law (as the Federal Trade Commission has been doing in the US). 

Meanwhile, the right wing UK government has effectively sided with the techno-optimists by leaving it to 90 regulatory authorities to try to assess the impact of AI in their sectors, and even cancelled plans for guidance on AI copyright licensing that copyright owners had requested. 

As the Finance Innovation Lab (of which I’m a Senior Fellow) has pointed out, the AI governance debate is dominated by those most likely to profit from more AI - and the voices of those who may be most negatively impacted are being ignored. Government needs to bring industry, researchers and civil society together, and find ways to include the perspectives of the wider public. To chart a safe route forward, it is essential that we prioritize the public interest, and align technology with societal values rather than the self-interest of the techno-optimists. 

Commercially speaking, however, there's also the point that consumers tend to reward businesses that act as 'facilitators' (who solve our problems) rather than 'institutions' (who solve their own problems at our expense). Of course, businesses can start out in one category and end up in another... The techno-optimists' commitment to their own self-interest (if recognised by consumers) should place them immediately in the second category.

Machine Unlearning: The Death Knell for Artificial General Intelligence?

Dall-E and toppng.com

Just as AI systems seem to be creating a world of their own through various 'hallucinations', Google has announced a competition between now and mid-September to help develop ways for AI systems to unlearn by removing "the influence of a specific subset of training examples — the "forget set" — from a trained model." This is key to allow individuals to exercise their rights 'to be forgotten', to object to processing, restrict processing or rectify errors under EU and UK privacy regulation, for example: Google accepts that in some cases it's possible to infer that an individual's personal data was used to train an AI model even after the personal data was deleted. But what does machine unlearning mean for the 'holy grail' of general artificial intelligence?

Unlearning is intended to be a cost effective alternative to completely retraining the AI model from scratch with the "forget set" removed from the training dataset. The idea is to remove  certain data and its 'influence' while retaining the accuracy or fairness of an AI model and its ability to generalize in ways that have already been held out as examples of what the model can achieve.

A problem with approaches to 'machine unlearning' to date has been inconsistency in the measures for evaluating their effectiveness, making comparisons impracticable. 

By standardizing the evaluation metrics Google hopes to identify the strengths and weaknesses of different algorithms and spark broader work on this aspect of AI.

As part of the challenge, Google will offer a set of information, some of which must be forgotten if unlearning is successful: the unlearned model should contain no traces of the forgotten examples, so that 'membership inference attacks' (MIAs) would be unable to infer that any of them was part of the original training dataset. 

Perhaps unlike the problem of hallucinations or fabrication (from which humans also suffer) - the advent of 'machine unlearning' provides another reason why 'artificial general intelligence' - a computer's ability to replicate human intelligence - will remain elusive, since humans often forget things only to recall them later, or are unable to recall events or aspects of them that we witnessed firsthand and/or were 'supposed' to remember (like an accident or a birthday or wedding anniversary).

Do You Know You Are Using And/Or Exposed To Artificial Intelligence?

Source: datamation

Amid the news that the Department of Work and Pensions is using artificial intelligence to decide benefits claims, a third of UK financial firms recently told the Bank of England and Financial Conduct Authority they were not using 'machine learning' (a subset of AI) in their businesses. That's pretty concerning, given the kind of AI we all use without realising and the fact that anyone wrongly denied benefits will need short term finance. That response also begs the question whether those firms know how their many stakeholders are using AI (whether unwittingly or not). If their suppliers, intermediaries, customers, investors, competitors and the government are using AI, then how does that affect their own strengths, weaknesses, opportunities and threats? And how does that in turn affect their stakeholders? No city on Earth is ready for the disruptive effects of artificial intelligence. Also worrying is the finding that smaller, fintech firms seem to believe that machine learning is no use to them. And given the problems with AI explained below, it's important for everyone to consider whether and how they rely on or are otherwise exposed to the two thirds of financial firms who are actually using AI... hyping the benefits without understanding the shortcomings will harm our trust AI where it could be helpful.

What is AI?

The term "AI" embraces a collection of technologies and applications, with machine learning usually featuring at some point:

• narrow artificial intelligence
• machine learning
• artificial neural networks 
• deep learning networks 
• automation
• robotics
• autonomous vehicles, aircraft and vessels
• image and facial recognition
• speech and acoustic recognition
• personalisation
• Big Data analytics

At the recent SCL event I chaired in Dublin, Professor Barry O'Sullivan explained that AI technologies themselves may be complex, but the concepts are simple. In essence, machine learning differs from traditional computer programming in that:

• traditionally, we load a software application and data into a computer, and run the data through the application to produce a result (e.g. how much I spent on coffee last year);

• while machine learning involves feeding the data and desired outputs into one or more computers or computing networks that are designed to write the programme (e.g. you feed in data on crimes/criminals and the output of whether those people re-offended, with the object of producing a programme that will predict whether a given person will re-offend). In this sense, data is used to ‘train’ the computer to write and adapt the programme, which constitutes the "artificial intelligence".

What is AI used for?

AI is used for:

1. Clustering: putting items of data into new groups (discovering patterns);
2. Classifying: putting a new observation into a pre-defined category based on a set of 'training data'; or
3. Predicting: assessing relationships among many factors to assess risk or potential relating to particular conditions (e.g. creditworthiness).

The Challenges with AI

The primary concerns about AI relate to:

1. cost/benefit - $50m in electricity to teach an AI to beat a human being at Go, hundreds of attempts to get a robot to do a backflip, but it can't open a door;
2. dependence on data quantity, quality, timeliness and availability;
3. lack of  understanding - AI is better at some tasks than humans (narrow AI) but general AI (same as humans) and superintelligence (better than humans at everything) are the stuff of science fiction. The AI that can predict 79% of European Court judgments doesn't know any law, it just counts how often words appear alone, in pairs or fours;
4. inaccuracy - no AI is 100% accurate;
5. lack of explainability - machine learning involves the computer adapting the programme in response to data, and it might react differently to the same data added later, based on what it has 'learned' in the meantime; 
6. the inability to remove both selection bias and prediction bias - adding a calibration layer to adjust the mean prediction only fixes the symptoms, not the cause, and makes the system dependent on both the prediction bias and calibration layer remaining up to date/aligned over time; 
7. the challenges associated with the reliability of evidence and how to resolve disputes arising from its use; and
8. there's a long list of legal issues, but lawyers aren't typically engaged in development and deployment.

This means the use of AI cannot be ignored. We have to be careful to understand the shortcomings and avoid hyping the benefits if we are to ensure trust in AI. That means challenging its use where the consequences of selection bias or false positives/negatives are fatal or otherwise unacceptable, such as denying fundamental rights or compensation for loss.

Instead, artificial neural networks and deep learning are better used to automate decision-making where "the level of accuracy only needs to be "tolerable" for commercial parties [who are] interested only in the financial consequences... than for individuals concerned with issues touching on fundamental rights."

Being realistic about AI and its shortcomings also has implications for how it is regulated. Rather than risk an effective ban on AI by regulating it according to the hype, regulation should instead focus on certifying AI's development and transparency in ways that enable us to understand its shortcomings to aid in our decision about where it can be developed and deployed appropriately.

Brits Look Away Now: Free Movement Of Non-Personal Data In the EU

The EU's "Digital Single Market" strategy has been boosted by an agreement to remove requirements for non-personal data to be stored in any one EU member state. The new law, approved in plenary by 520 votes to 81, with six abstentions, is due to be approved by the EU Council of Ministers on 6 November. It will apply six months after its publication in the EU Official Journal.

Restrictions on the free movement of personal data have long been relaxed under the EU data protection framework. The latest move is expected to double the value of the EU data economy to 4% of GDP by 2020.

The agreement builds on a report into "Building a European Data Economy",  statistics on the use of cloud computing by businesses, a study on the legal and administrative restrictions on the free flow of non-personal data (often mistaken) and study of the economic benefits of cloud computing uptake and free flow of data within the EU.

In summary, the agreement means that:

• public authorities cannot impose "unjustified" data localisation restrictions;

• the data remains accessible for regulatory and supervisory control even when stored or processed across borders within the EU;

• cloud service providers are encouraged to develop self-regulatory codes of conduct for easier switching of provider and porting data back to in-house servers, by mid-2020;

• Security requirements on data storage and processing remain applicable when businesses store or process data in another Member State or outsource data processing to cloud service providers;

• Single points of contact in each Member State will liaise with other Member States’ and the Commission to ensure the effective application of the new rules.