AI: What is the future of artificial intelligence? - BBC News

Is artificial intelligence (AI) an opportunity, a threat, or even both?

Artificial intelligence is a branch of computer science which develops machines and software which can perform tasks which normally require humans to do them, for example, decision-making or speech recognition.

Experts are divided on how this technology should be regulated, and who by.

#ArtificialIntelligence #AI #BBCNews

[deeliciousplum]

15:39 "In order to respond to it (the future of AI as well as any concerns) effectively, we've got to start with education. And, right know, I mean, I've tried to brief policy makers on this. It's like explaining particle physics to a chocolate chip cookie."
Priceless and helps us to understand the challenges that people who have legitimate concerns have with raising them to those in positions of influence and power.

[aether]

@15:45 "I've tried to brief policy makers on this and it's like trying to explain particle physics to a chocolate chip cookie" Absolutely roasted them with this one lol, rarest insult I've seen in a minute 😂

[kaziloser]

I enjoyed this convo as it was talking about pros and cons of AI and how we can measure it.

[jonnyhatter]

On the actual technical development side of this, I kinda don't see how we can avoid dangers or put in guardrails when the engineers themselves have shown that large language models have emergent properties. Emergent meaning sudden behavior that the LLM wasn't trained on. So whatever you train it on doing or not doing, you can never account for an emergent behavior that might curtail a safety protocol or something. In other words, ultimately it would be impossible to contain this technology safely. Guys, GPT 4 is still a baby technology. Just two generations before it, GPT-2 totally freaking sucked. In just two generations, this technology has gotten ridiculously good.

[robynwyrick]

I am generally a techno optimist, but I think this was an exceptional panel. Great work folks. Yes, get policy makers on the ball.

[zoeherriot]

The guy saying we should be thinking about what three wishes we should ask the genie, has obviously never read a story about how that turns out for the person making the wishes.

[SeymourClevage]

They can't even regulate the internet, Ai won't be any different.

[Olmirz]

Definitely the most important BBC interview ever done, and hopefully, it'll become the most consequential 🤞

[EC]

6:25 - Not sure if we're very far from AGI..look at GPT 4 (or even 5 relatively soon to come next) and AutoGPT ..it already is feeling like the beginning of a basic primitive AGI..give this like 2-3 more years and we'll almost certainly end up with some form of AGI, the progress of AI lately is going quickly within months or even weeks there are new developments.

[SmokeyMcPot]

That guy said it perfectly. Even if AI doesn't get any better (which it 1000 percent will) it will already profoundly change the world as we know it. Exciting / scary time to be alive.

[jianhaoqiu]

I think these panelists are absolutely right. Each kind of technology is a double-edge sword. We should get it under human control before it is too late. But the reality is always disappointing. Just like tackling climate change, which has been a slogan for so many years but very little concrete actions have been taken and the situation is still getting worse. So I assume that the same may happen to AI. Everyone knows what is the right thing to do, to impose restrictions before all regulations including ethics are in place, but no one is gonna do it until real risk comes...

[johnhoy]

Well done BBC for a balanced and interesting panel. These people actually appeared normal and not hyperbolic

[patrickcollins]

Nothing can halt the march of progress. We are about to witness the beginning of a glorious age in human history.

[nmarks]

The man most directly responsible
is Miles Bennett Dyson, director of special projects at Cyberdyne Systems Corporation.

[DrBettyQuantumAI]

The same challenges are experienced on business in general, including SMEs. Thank you for bringing this topic to a broader discussion. AI marketing is the future, but we must understand that it is far more than just Chat GPT.

[andrewedgar]

Got to hand it to Miles Taylor for these brilliantly creative analogies 😂

[AnonymousBosch]

China and US are in a competition over AI, someone who thinks that is possible a slow down in the development of this technology is just deeply naive.

[jasonwang]

Here is my suggestion on how to regulate AI, which is to require or demand establishment of AIAL (standing for AI Activity Log) for every AI network, which is similar to the user activity log of a social media (user) account or any types of network (user) account.
For example, every content, product, or work produced by an AI shall have an content ID or the like (indicating it's produced by an AI), and such content ID shall be included or recorded in the AIAL; other info (such as brief description of the content, owner of AI network, etc) may also be included therein.
The AIAL format should be standardized.

[spadhnik]

Great panel and thought provoking discussions

[yukokeller]

As an educator and parent, my primary concerned is that there may be an even greater difference in student academic achievement than there is today.