🚨 World’s major powers sign AI treaty

ALSO: Google’s new AI photo features

Estimated Read Time: 3 minutes

AI is the biggest scare most people have right now. Aside from the obvious job concerns is the hanging question… how do we stop AI from becoming a literal threat to humanity? The biggest powers of the West think they have a solution and have signed the biggest global AI treaty we’ve seen so far to create a standard for AI safety measures.

• 📝 The AI Safety Treaty has been signed.

• 🪖 South Korea’s military AI summit.

• 📸 Google’s new AI photo features.

• 📈 Reddit acquires AI startup to boost their ad business.

📝 The AI Safety Treaty has been signed

What happened: Several major countries, including the US, UK, and EU, have signed a groundbreaking international AI safety treaty. This legally binding agreement aims to ensure AI systems respect human rights, democracy, and the rule of law. Let’s check it out.

The details: 

• The treaty is called the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy, and the Rule of Law. Quite a mouthful.

• What makes the treaty special? It’s the first legally binding global treaty on AI safety. AI companies in these countries will need to comply.

• The treaty requires countries to set up regulatory bodies to manage AI risks and applies to all forms of AI. The problem here will, as always, be execution.

• This is separate from the EU’s AI Act and the UK’s AI Safety Summit initiatives and even complements them.

• Some of the rules in the treaty include non-discrimination, transparency, accountability, required human oversight and a right to privacy.

Takeaway: It’s about time AI gets some regulations. But the biggest problem here are not the rules but rather the execution. How these regulations are implemented will be the deciding factor in their success.

🤖 Elon Musk insists xAI has no reason to license their models to Tesla. Meanwhile shareholders in Tesla are suing Musk over starting xAI as they believe it’s a “competing company”.

📆 ZipRecruiter’s new tool makes scheduling calls easier. The new AI tool is called ZipIntro, which allows employers to schedule and conduct quick introductory calls with job candidates. The tool matches candidates based on key job criteria and helps employers connect within minutes of posting a job.

🪖 South Korea is holding an AI summit to standardize AI’s use in military tech. Over 90 countries, including the U.S. and China, sent representatives to work on a non-binding blueprint for AI’s military application. Ukraine’s use of AI-powered drones in the Russia-Ukraine war was highlighted as an example of AI’s potential in warfare. I fully expect none of these countries to abide by any limitation that would prevent them from gaining an edge.

🇺🇸 OpenAI and Anthropic will share their models with the US government. This allows the US government access to their AI models both before and after public release. None of this is a surprise considering how Microsoft and Meta also have various ties and deals with the US government.

📸 Google’s new AI photo features

What happened: Google’s introducing an AI-powered “Ask Photos” feature to their search. Now you can search for photos, including those in your private collection, using natural language queries.

The details:

• Google’s Ask Photos feature is available in early access for US users.

• Ask Photos allows users to search for photos using natural language. 

• EXAMPLE: In the example photo at the top, the user asks what they ate at a certain place. Google finds the photo and understands the food in the picture. Pretty cool.

• Google names another example as, “find the best photo from each of the National Parks I visited.”

• You can also ask questions about the photos, such as what themes you used for your children’s previous birthdays.

• Google also announced their Google Shopping try-on tool will now support dresses, which were apparently the most searched-for clothes in the tool. They’re also the hardest to implement, as the dresses are more complex than other clothing.

• The try-on tool allows you to, you guessed it, try on clothes virtually using AI before a purchase. Pretty soon all clothing brands will use something like this. I image these features will eventually integrate naturally into social media apps and ads.

Why it matters: These two features will potentially become major features we use in our everyday lives. Imagine a future where social media knows what your virtual avatar looks like, and every ad you see shows your virtual avatar wearing those clothes or using those products. Someday, maybe. The biggest current concern is simply privacy.

📈 Reddit acquires AI startup to boost their ad business

What happened: Reddit has acquired Memorable AI, a startup that uses AI to predict which ads will attract attention and boost engagement.

The details:

• Memorable AI uses machine learning to predict which images and videos will capture users’ attention. Maybe someday split-testing ads will be a thing of the past?

• Reddit plans to integrate this machine learning tech into their ad platform to improve ad efficiency.

• Memorable’s AI can automatically tweak ad design, such as changing logo size or text color, to improve performance. That being said, these are small factors in an ad’s success compared to the creative content itself.

• Brands like Unilever and L’Oréal already use Memorable’s AI to enhance their ads.

• This acquisition follows Reddit’s 2022 purchase of AI startup Spiketrap, which improved ad targeting based on content sentiment. Obviously they’re going big with ads.

• Reddit recently went public after using access to their platform to strike several deals with big AI companies who want to use Reddit as “training data”.

• The deal is reportedly worth $40 million, a small price to pay if it improves Reddit’s ad revenue.

• As suspected, the response from Reddit users is not positive.

Takeaway: AI has infinite potential to improve ad efficiency and how we utilize brands. We run ads ourselves and know AI could make things dramatically better… if they do it right. But users on Reddit are not happy. They suspect they’ll get even more ads or bots trying to start positive conversations about brands. And yeah, I agree. Tech like this will result in more ads. Yay?

⚡️ If the Greeks uses AI…

Source: Midjourney

Give us your feedback!

Got feedback? Rate today’s newsletter by clicking below! 

Got more personal feedback? Fill in the box when you rate us and give us your criticism and feedback.

Thank you for reading! 

❤️Share the (ARTIFICIAL) Love!

Got a friend who wants to appear as smart as you? An AI fanatic or someone shaking in their boots of getting replaced? Share us by clicking the link below and you’ll get our love ;)