AI news recap: Biased driverless cars; prompt engineers; chip shortages; end of CAPTCHA

What is an ‘AI prompt engineer’ and does each firm want one?

Artificial intelligence is succesful of superb feats, from writing a novel to creating photorealistic artwork, however evidently it isn’t so good at extracting precisely what we would like. It fails to know nuance or overcome poorly worded directions. That has given rise to the brand new job of “prompt engineer” – people who find themselves expert at crafting the exact textual content directions wanted for AI to supply precisely what is required – typically with salaries of upwards of $375,000 a 12 months.

This skill to unlock the potential of AI with their “magic voodoo” could appear to be a bit of a fad, however New Scientist discovered that tons of corporations discover it surprisingly useful – in the mean time, no less than. The query is whether or not AI will turn into higher at understanding what people imply and due to this fact minimize out the intermediaries.

Driverless vehicles could battle to identify youngsters and dark-skinned individuals

Racial bias in AI is nothing new, and sadly it’s a trait inherited from knowledge tarnished by human prejudice. Previous instances have made it tougher for these with darkish pores and skin to get a passport or be shortlisted for a job, but it surely has now emerged that AI may additionally place them at larger threat of being hit by a driverless car.

AI of the sort utilized in driverless vehicles is 7.5 per cent higher at recognizing pedestrians with mild pores and skin than these with darkish pores and skin, warn researchers. Part of the issue is the shortage of pictures of dark-skinned pedestrians in coaching knowledge. Racial bias of all kinds must be rooted out of AI, however when it’s probably life-threatening it’s crucial to behave swiftly earlier than that expertise is launched into the true world.

Bots are higher at beating ‘are you a robot?’ checks than people are

The Turing check is a well-known proposal for distinguishing AI from people, however in phrases of scale there was no larger check than the widespread use of CAPTCHA – the irritating little issues it’s important to resolve when signing as much as varied web sites. Whether it’s clicking the tiles that embrace visitors lights, typing in distorted textual content or fixing an arithmetic downside, the concept is identical: to permit people previous whereas stopping AI bots whose intention is to abuse the location.

The downside is that AI has turn into higher than people at these checks. Rather a lot higher. More correct and quicker. It appears that each one CAPTCHA checks are managing to do is irritate people. So is it time to ditch them altogether? And ought to we be apprehensive that engineers battle to give you an on-screen activity that people can do higher than AI?

Chip shortages are producing winners and losers within the AI gold rush

It is not any secret that AI is massive enterprise in the mean time. The arrival of generative fashions has birthed a technology of startups and each massive firm is dashing to construct, borrow or purchase a mannequin to streamline some half of their enterprise. The result’s that the {hardware} generally used to coach and run these fashions – initially designed to energy pc video games – is in extraordinarily brief provide.

Builders of these chips are making hay whereas the solar shines as new gamers are understandably muscling in. Meanwhile, commerce sanctions are making the worldwide provide of chips a political situation, and academia is more and more priced out of the AI analysis that it kickstarted. AI could really feel like a software program revolution, however additionally it is very a lot a {hardware} arms race.

AI chatbots turn into extra sycophantic as they get extra superior

If you’re in search of straight solutions from AI chatbots you may need an issue: they appear to simply inform us what we wish to hear. These digital yes-men are “designed to fool us and to kind of seduce us”, says Carissa Véliz on the University of Oxford. The downside appears to develop worse as the scale of the mannequin will increase, which is a severe situation as a result of rising scale at the moment appears to be the easiest way to make them extra succesful.

How can we belief an AI if it responds to our questions not with info or proof, however with a re-jigged reflection of our personal opinions and biases? And ought to we actually be including that expertise to search engines like google and yahoo earlier than we now have discovered an answer?