Images are unavailable offline.

The recent hype over AI is much like the same fever that had fuelled crypto, when once you could slap 'blockchain' on the name of any company and see its stock soar fourfold.

Martin Meissner/The Associated Press

Amid all the hoo-ha over artificial intelligence this year, Microsoft Corp. MSFT-Q, which has a stake in the laboratory behind the ChatGPT bot, has seen its shares go up more than 25 per cent.

Various AI stocks, with names youve never heard of, are hotter than hot, even with a recession looming and at the foot of a tech beatdown in the markets. BigBear.ai Holdings Inc., an information-technology services company, is up about 250 per cent on the year; at one point in February, it was up 700 per cent.

Wanna make money? Boy, do I have a great idea for you. Just add AI to the name of your company. Theres a voice-recognition company that used to be called SoundHound Inc., but went public in 2022 as SoundHound AI Inc. SOUN-Q. The stock has admittedly pared back some gains since then, but it is still up nearly 100 per cent for the year.

Story continues below advertisement

Any of this sound familiar? Its the same fever that had fuelled crypto, when once you could slap blockchain on the name of any company and see its stock soar fourfold. Im pretty sure that soon, as with crypto, the term AI bro will enter the lexicon to describe a young man who is passionate and enthusiastic about the industry.

Oh, wait it has. An Urban Dictionary entry for AI bro was made in January of this year.

Will AI take over the world? And other questions Canadians are asking Google about the technology

Lets face it, AI is the new crypto. All the hype, investment mania and scams of past years investment cycles are going to come back.

To that, you might slam your table, squint your eye around your monocle and say: Wait thats not right! At least AI does something. Crypto is just make-believe money!

Story continues below advertisement

A commonly expressed view. And a wrong one. But lets for the sake of argument say that it is correct. Has that distinction resulted in any difference in the markets?

It wasnt just 2020, the year of the really expensive digital pictures, or NFTs, that crypto was booming. Remember 2017, when a market frenzy was sparked by the Canadian-founded Ethereum, which let anyone easily create their own coin?

At one point that year, the furniture chain Ethan Allen Interiors Inc. ETD-N was up 50 per cent, largely attributed to how its ticker at the time, ETH, was the same as the abbreviation for Ethereums ether coin.

While Ethan Allen eventually changed its ticker to distance itself, others fiercely coveted that nominal crypto association.

Story continues below advertisement

I wasnt being hyperbolic when I wrote earlier that companies can slap blockchain onto their names and see their stock quadruple in value. That was exactly what happened when Long Island Iced Tea Corp. changed its named to Long Blockchain Corp.

Meanwhile, Eastman Kodak Co. KODK-N, the camera maker, saw its stock triple in value after a bad year by announcing it would go into crypto mining.

Then there were the outright scams. The infamous OneCoin raised US$4-billion, but there is no evidence it had even developed a digital currency based on blockchain technology. Such scams are so plentiful that the U.S. Justice Department is still announcing new 2017-era cases to this day.

Such scams abounded because they were easy. Regardless of what many think of it, there are defined metrics for what makes a cryptocurrency namely in terms of the code that goes into it. But people cant see or hold a cryptocurrency. So, its easy to claim youve made one. The end user doesnt always have the sophistication to tell the difference until its too late.

Story continues below advertisement

Again sounds familiar? Have you ever wondered how many purported AI projects are actually AI?

A London-based startup, Engineer.ai, once claimed to use artificial intelligence to help people build apps. It attracted US$30-million from investors, including a unit of Japans SoftBank Group Corp. SFTBY. The Wall Street Journal later reported that Engineer.ais AI claims were greatly exaggerated actual humans in India were building the apps.

Such practices are so rampant, there is even a neologism coined for it: AI washing.

What it all boils down to is this: When crypto entered the mainstream, it was hard to define or even understand. In that messy environment, companies thrived and empires were built and so also rose the scams and OneCoins of the world. AI is having the same moment now.

Follow this link:
Opinion: Let's face it, artificial intelligence is becoming the new ... - The Globe and Mail

Law firms have been racing to adopt artificial intelligence after developments in the technology have enabled it to draw up contracts, assist due diligence processes and draft legal opinions.

The launch of natural language chatbot ChatGPT in November marked a significant turning point in generative AI. Created by Microsoft-backed OpenAI, the bot produces convincing and humanlike sentences, using large language models to predict the likely next word in a sequence.

It has led other big tech companies, including Google and Microsoft itself, to quickly follow suit. Start-ups have also been leveraging the underlying technology used in these products to develop specialist AI for legal services.

Law firms and consultancies are now using the software to automate tasks and drive efficiencies, spurred to cut costs by falling revenues amid a corporate dealmaking drought. Magic Circle law firms and Big Four accounting groups have been experimenting with AI platforms built for legal tasks such as drafting contracts, translating documents into different languages, and suggesting legal opinions.

Recommended

We get several thousand queries daily, and it is pretty consistent across the offices...in multiple languages [and] different areas of law, says David Wakeling, head of Allen & Overys markets innovation group, which comprises both lawyers and developers. The law firm was the first Magic Circle firm to adopt generative AI and has been using an eponymous product by a US start-up called Harvey AI since November. It is [now] a serious part of the operating model, he explains. We are way past trial.

Harvey was built using the GPT language models created by OpenAI, which has also invested in the US start-up. As well as the general internet data that underlies GPT, Harvey is trained in legal data including case law. The system alerts A&Os lawyers to fact-check the content it creates, as generative AI is known for hallucinating: stating things confidently as fact, despite there being no basis for them in reality.

It is a blank page, quick first stab, Wakeling says. You know you are always going to edit it; it is never good enough. But [if] you apply that to 3,500 people, that is a serious saving in terms of time; an hour or two a week is a big deal.

We need human lawyers to ensure we keep expanding and pushing forwards the law so it doesnt atrophy

Almost half of all current tasks in the legal profession could be replaced by AI, according to a recent report by Goldman Sachs. Automating them would eliminate the need for humans to carry out some of the more administrative and mundane work although it could also mean that trainee solicitors and graduates at law firms no longer got to experience it.

We need to think about how we train young lawyers, says Kay Firth-Butterfield, head of artificial intelligence at the World Economic Forum. They cannot all instantly be able to advise on the most complex matters or do complex client meetings, or think of ways to challenge the status quo. This has to be learned.

She sees limits to AIs use: Because all that generative AI can do is look at historical data to give answers, we need human lawyers to ensure we keep expanding and pushing forwards the law so it doesnt atrophy.

Protagonists argue that AI-based tools will free up lawyers to do more skilled work and give strategic advice, while saving time and reducing costs for firms and clients.

It definitely reduces the billable hours, says Richard Robinson, founder and chief executive of Robin AI, which launched in 2019 and provides AI-based legal software. But he points out: The best firms want to be paid for high-level strategic work, things that fundamentally, at least today, no AI is trying to replicate like high-level negotiations, insights into whats happened in other [similar] deals in the market.

$1bnPwCs investment in AI to automate parts of its audit, tax and consulting business over the next three years

Robin AI works with two of the Big Four accounting firms, as well as private equity funds and law firm Clifford Chance. It sells software and also offers an added service, where its team of 30 in-house lawyers and paralegals oversee the AI-generated results. Its technology can also be used to scan legal documents to assess risk exposure.

But Robinson warns that these tools are basically not ready to be used without people safeguarding them highlighting that the pure output of such technologies should always be checked and edited by a qualified expert.

Large language models improve as more data is fed into them and as they are put to more use. PwC, which uses Harvey for mergers and acquisitions, due diligence and drafting contracts, says it has had an influx of clients saying they want to adopt AI, but are concerned about data protection.

Data confidentiality and security is paramount and really important...because data is sensitive and there is legal privilege, says Sandeep Agrawal, a partner in PwCs tax and legal services. Agrawal met executives from Harvey recently to discuss how it can ringfence data and encrypt the information, meaning it is more secure. Harvey segregates all customer data and offers encryption tools to protect access to client information.

Recommended

In a sign of growing confidence in the technology, PwC last week pledged to invest $1bn in AI to automate parts of its audit, tax and consulting business over the next three years.

Jerry Ting, chief executive of contract management business Evisort and a lecturer at Harvard Law School, reports a similar shift towards adopting AI in recent months. Evisort, which launched in the US in 2016, offers AI software that allows clients to create and manage contracts including drafting and signing through an automated process.

Before GPT, it was: here is what AI is, heres why it benefits you; it was almost that we had to convince them, he says. Now, they are showing up at the door already convinced. The question becomes: how do I use it in a way thats safe, that actually drives my business outcomes, and it does fit in my budget?

This article has been updated to reflect that Evisort is a contract management business

Original post:
Law firms embrace the efficiencies of artificial intelligence - Financial Times

In todays commentary I delve into the question of open-source versus closed-source models for AI, and how this will shape the future of the internet.

It all started with leaked Google documents shared by an anonymous insider to

But the uncomfortable truth is, we arent positioned to win this arms race and neither is OpenAI. While weve been squabbling, a third faction has been quietly eating our lunch.

Im talking, of course, about open source. Plainly put, they are lapping us.

Ive been thinking about this open source since Stable Diffusion came out last summer. But the last two months, since Metas LLaMa model was leaked has seen a rich ecosystem of developers swarm around open-source LLMs. And things are moving quickly.

But before I share my thoughts, Ill first summarise the key observations the Google insider makes.

The first observation is that the gap between the current state of the art (GPT-4) and open-source models is changing quickly. For example, a $100 open-source model with 13 billion parameters is competing with a $10 million Google high-end model with 540 billion parameters.

Secondly, the open-source community has solved many scaling problems through a range of optimizations. MosaicML is a great example of this, demonstrating that they can train a Stable Diffusion model, which is not a large language model, six times cheaper than the original.

The third observation is that the dynamic of the open-source market creates a faster rate of iteration. This is because there are many developers who are contributing to the market, leading to a much faster rate of learning. Learning is a key factor in the Exponential Age; it drives down cost and improves price performance. Potentially, open-source models could learn and iterate far faster than closed-source ones.

In the last few days, Ive spoken to Emad Mostaque from StabilityAI, Yann LeCun from Meta, and several other people who are developing these things. In addition, Ive been thinking about open-source and public goods for more than two decades. Heres my take

Original post:
The open-source future of artificial intelligence - Exponential View

Open this photo in gallery:

Demonstrators hold signs during the 2007-2008 Writers Guild of America strike in Hollywood.CHRIS DELMAS/AFP/Getty Images

Gus Carlson is a U.S.-based columnist for The Globe and Mail.

For decades, Hollywood writers have been creating stories about a future where machines take over the world. Think The Terminator, Blade Runner, I, Robot even the animated kids movie, WALL-E.

Now, these creatives find themselves on the thin edge of the wedge in their own version of that apocalyptic plot line.

When the members of the Writers Guild of America went on strike this week, they listed among their demands a provision that nods to a not-too-distant future where human creativity is under siege: regulations for the use of materials produced using artificial intelligence or similar technologies.

Beyond the cruel irony of this existential crisis for writers, the call to limit AIs influence in this context raises the question: Does it really matter who or what creates a good story well told?

Would viewers really care if their favourite Netflix series was the product of AI, as long as it was engaging and entertaining and especially if they didnt have to wait so long between seasons?

Purists would say they should care that the human creative process is iterative and by nature takes time to brew. Great art, whether it is writing, music, film, stage, painting, dance or sculpture, is about the expression of human emotion and feeling that cant be captured and replicated by machines. AI can do many things to mimic art, but is it really art?

Sure, AI can write a plot line about racism in Depression-era Alabama, but can it capture the powerful anxiety of Harper Lees To Kill and Mockingbird? It can spew out a scene where two friends kibitz about love in a New York diner, but can it capture the comic brilliance of a line of dialogue like Ill have what shes having? And yes, it can mimic a complicated guitar solo, but can it inspire like the magic of a B.B. King free-form riff?

Increasingly, however, the consumers and producers of content might not be so quick to dismiss the idea of tech-driven shortcuts to feed our instant-gratification culture.

As production costs for films, television programs and streaming series rise, and the demand to fill the content pipeline intensifies, the use of AI is becoming a real option for studios and networks. If they can produce high-quality content faster and cheaper and the viewers and subscribers dont really care or cant even tell how the sausage is made everybody, except the writers, of course, wins.

That point is more salient when we consider that restricting AI in the creation of stories isnt the only thing the writers want as part of their contract negotiations with a trade association representing the top Hollywood studios, television networks and streaming platforms.

They have many more concerns in the here and now, including making more money. A big part of their gripes is that streaming series typically have fewer episodes than broadcast shows, so maintaining a consistent income stream is difficult.

These demands will further strain the budgets of the studios. If writers want to be paid more, AI would start to look all the more attractive to the studios.

Of course, Hollywood writers are not alone in their wariness of a creative world infected by technology. Book publishers are on the lookout for AI-manufactured manuscripts, and many college admissions officers are placing less weight on student essays some are eliminating them as a requirement altogether because of widespread use of AI to create personal stories so expertly written they could not have come from the keyboard of the average teenager.

As writers and other artists struggle to protect their gifts, the broader cultural challenge is clear. There are many things AI can do, as well as many things it cant. The quandary for creatives is whether the difference will continue to matter to the average humanoid consumer of their wares. The economic viability of their craft hangs in the balance.

Read this article:
Opinion: Striking television and film writers want artificial intelligence ... - The Globe and Mail

Artificial intelligence (AI)

UKs outgoing chief scientist urges ministers to get ahead of profound social and economic changes

The new genre of AI could be as transformative as the Industrial Revolution, the governments outgoing chief scientific adviser has said, as he urged Britain to act immediately to prevent huge numbers of people becoming jobless.

Sir Patrick Vallance, who stood down from his advisory role last month, said government should get ahead of the profound social and economic changes that ChatGPT-style, generative AI could usher in.

However, in a wide-ranging final parliamentary hearing that also covered his reflections on the pandemic and the rise of China as a global scientific power, he suggested AI could also have considerable benefits that should not be overlooked.

There will be a big impact on jobs and that impact could be as big as the Industrial Revolution was, Vallance told the Commons science, innovation and technology committee. There will be jobs that can be done by AI, which can either mean a lot of people dont have a job, or a lot of people have jobs that only a human could do.

In the Industrial Revolution the initial effect was a decrease in economic output as people realigned in terms of what the jobs were and then a benefit, he added. We need to get ahead of that.

Vallance called for a national review of which sectors would be most significantly affected so plans could be drawn up to retrain and give people their time back to do [their jobs] differently.

The comments follow an announcement by IBM this week that it is suspending or reducing hiring in jobs such as human resources, with a suggestion that 30% of its back-office roles could be replaced by AI in five years.

Echoing comments by the AI pioneer Geoffrey Hinton, who announced his departure from Google this week, Vallance said the most immediate concern posed by AI was ensuring it did not distort the perception of truth.

He added that there was also a broader question of managing the risk of what happens with these things when they start to do things that you really didnt expect.

Despite these potential existential threats, the technology also presented opportunities, Vallance argued. In medicine, it could be that you get more time with your doctor rather than being pressurised, he said. That could be a good outcome.

We shouldnt view this as all risk, he added. Its already doing amazing things in terms of being able to make medical imaging better. It will make life easier in all sorts of aspects of every day work, in the legal profession. This is going to be incredible important and beneficial.

Vallance, who is now chair of the Natural History Museum, appeared sceptical about the prospect of developing a British version of ChatGPT, dubbed Brit-GPT, which some experts have called for in recent months. In March, the Treasury committed 900m to building a supercomputer to boost sovereign capabilities in this area.

Vallance said the focus for the UKs core national capability should be on understanding the implications of AI models and testing the outputs not on building our own version.

He said: You need to be able to probe them and understand them. I just dont think the idea were going to invent something that rivals what the big companies have already made is very sensible. It sounds like attempts to invent a new internet. I mean, why?

Vallance also implied that a moratorium on AI would not be feasible. Unilaterally falling behind doesnt seem to me to be a sensible approach, he said.

Looking back over his tenure, Vallance said his proudest achievements included helping establish the Covid-19 vaccines taskforce and acting as chief scientific adviser for the Cop26 climate summit.

He said he regretted very clumsy wording about herd immunity that led to misunderstanding and controversy early in the pandemic. In a March 2020 interview, Vallance said the aim was not to suppress the virus completely to build up some degree of herd immunity whilst protecting the most vulnerable.

He told the committee his intention was to reflect that immunity was fundamentally how you end pandemics rather than it being an intended strategy. People get immunity through vaccines and they get immunity through catching infections, he said. Ultimately that is where we have got to.

On the origins of the pandemic, Vallance said by far the most likely explanation was a spillover from bats, and that the available evidence suggested a lab leak was less likely.

Vallance also commented on the UKs position in a shifting geopolitical world, with countries including China in the ascendancy in science and technology. Against this backdrop, he said, it was essential for the UK to remain part of the EUs Horizon programme, pointing out it took the flagship research scheme a decade to get going effectively.

The idea that you can instantly set up something equivalent is flawed, he said. China has huge scale, US has huge scale. There are some parts of science that need scale. You cant replicate that domestically.

He called on the UK government to make changes to its visa scheme, which he said needed to be quick and internationally competitive in order to attract the best scientists. When asked whether the Home Office had responded to his advice on this, Vallance said: I guess the feedback is the action.

{{topLeft}}

{{bottomLeft}}

{{topRight}}

{{bottomRight}}

{{.}}

Read the rest here:
AI could be as transformative as Industrial Revolution - The Guardian