AI news recap for July: While Hollywood strikes, is ChatGPT getting worse?

Hollywood actors strike over use of AI in movies and different points

Artificial intelligence can now create photographs, novels and supply code from scratch. Except it isn’t actually from scratch, as a result of an enormous quantity of human-generated examples are wanted to coach these AI fashions – one thing that has angered artists, programmers and writers and led to a collection of lawsuits.

Hollywood actors are the newest group of creatives to show in opposition to AI. They concern that movie studios may take management of their likeness and have them “star” in movies with out ever being on set, maybe taking up roles they’d quite keep away from and uttering strains or performing out scenes they’d discover distasteful. Worse nonetheless, they won’t receives a commission for it.

That is why the Screen Actors Guild and the American Federation of Television and Radio Artists (SAG-AFTRA) – which has 160,000 members – is on strike till it might probably negotiate AI rights with the studios.

At the identical time, Netflix has come underneath hearth from actors for a job itemizing for individuals with expertise in AI, paying a wage as much as $900,000.

AIs educated on AI-generated photographs produce glitches and blurs

Speaking of coaching knowledge, we wrote final yr that the proliferation of AI-generated photographs may very well be an issue in the event that they ended up on-line in nice numbers, as new AI fashions would hoover them as much as prepare on. Experts warned that the top consequence can be worsening high quality. At the danger of constructing an outdated reference, AI would slowly destroy itself, like a degraded photocopy of a photocopy of a photocopy.

Well, fast-forward a yr and that appears to be exactly what is occurring, main one other group of researchers to make the identical warning. A staff at Rice University in Texas discovered proof that AI-generated photographs making their approach into coaching knowledge in giant numbers slowly distorted the output. But there is hope: the researchers found that if the quantity of these photographs was saved beneath a sure degree, then this degradation may very well be staved off.

Is ChatGPT getting worse at maths issues?

Corrupted coaching knowledge is only one approach that AI can begin to collapse. One examine this month claimed that ChatGPT was getting worse at arithmetic issues. When requested to verify if 500 numbers had been prime, the model of GPT-4 launched in March scored 98 per cent accuracy, however a model launched in June scored simply 2.4 per cent. Strangely, by comparability, GPT-3.5’s accuracy appeared to leap from simply 7.4 per cent in March to virtually 87 per cent in June.

Arvind Narayanan at Princeton University, who in one other examine discovered different altering efficiency ranges, places the issue all the way down to “an unintended side effect of fine-tuning”. Basically, the creators of those fashions are tweaking them to make the outputs extra dependable, correct or – doubtlessly – much less computationally intensive as a way to minimize prices. And though this may increasingly enhance some issues, different duties would possibly endure. The upshot is that, whereas AI would possibly do one thing properly now, a future model would possibly carry out considerably worse, and it might not be apparent why.

Using larger AI coaching knowledge units might produce extra racist outcomes

It is an open secret that a variety of the advances in AI in recent times have merely come from scale: bigger fashions, extra coaching knowledge and extra laptop energy. This has made AIs costly, unwieldy and hungry for assets, however has additionally made them way more succesful.

Certainly, there is a variety of analysis happening to shrink AI fashions and make them extra environment friendly, in addition to work on extra swish strategies to advance the sphere. But scale has been a giant a part of the sport.

Now although, there is proof that this might have severe downsides, together with making fashions much more racist. Researchers ran experiments on two open-source knowledge units: one contained 400 million samples and the opposite had 2 billion. They discovered that fashions educated on the bigger knowledge set had been greater than twice as more likely to affiliate Black feminine faces with a “criminal” class and 5 occasions extra more likely to affiliate Black male faces with being “criminal”.

Drones with AI focusing on system claimed to be ‘better than human’

Earlier this yr we coated the unusual story of the AI-powered drone that “killed” its operator to get to its meant goal, which was full nonsense. The story was shortly denied by the US Air Force, which did little to cease it being reported all over the world regardless.

Now, we’ve contemporary claims that AI fashions can do a greater job of figuring out targets than people – though the main points are too secretive to disclose, and subsequently confirm.

“It can check whether people are wearing a particular type of uniform, if they are carrying weapons and whether they are giving signs of surrendering,” says a spokesperson for the corporate behind the software program. Let’s hope they’re proper and that AI could make a greater job of waging battle than it might probably figuring out prime numbers.

If you loved this AI news recap, strive our particular collection the place we discover probably the most urgent questions on synthetic intelligence. Find all of them right here:

How does ChatGPT work? | What generative AI actually means for the financial system | The actual dangers posed by AI | How to make use of AI to make your life easier | The scientific challenges AI is serving to to crack | Can AI ever turn out to be aware?