🏷️ Meta to label AI-generated content

💼 Work in AI GPT By AI Tool Report

AuthorAuthor

Martin Crowley & Liam Lawson
February 07, 2024

This morning’s top story: Meta to start labeling AI-generated images to stop unsolicited fakery.

We’re launching our new “Work In AI” job board on Monday (sign up for early access) and have developed a series of Work In AI GPTs, designed to help you land your dream job.

Today, we’re releasing edition 2 of 3: The Practice Interview. Use it to practice key interview questions ahead of the big day.

💼💼💼💼

Plus, we’re celebrating! We’ve just reached 500,000 subscribers, which is a major milestone for us, and we’d like to thank all our subscribers—some of whom have been with us since day 1—for your continued support in helping us make the AI Tool Report the biggest AI newsletter in the world.

We’ll continue to bring you the most comprehensive AI coverage in 5 minutes, daily…

🥳🥳🥳🥳

🌤️ This Morning on AI Tool Report

  1. 🏷️ Meta to label AI-generated content

  2. ☎️ Biden’s robocall perpetrators revealed

  3. 🥼 OpenAI leads $70M funding for AI healthcare assistant

  4. 🛟 Google and Meta clash with UK over AI safety

  5. 🛒 IKEA releases new AI shopping assistant

Read Time: 5 minutes

FINANCE

Brought to you by YellowBrick

AI REGULATIONS

Our Report: Meta is building an AI-powered tool that will automatically detect and label (in all languages) images—posted on Facebook, Instagram, or Threads—that have been generated using AI.

🔑 Key Points:

  • Meta already applies watermarks to images created with its own ‘Imagine with Meta’ AI image generator to establish whether the images are real or fake.

  • Its new tool will use invisible watermarks and embedded metadata to identify and label AI images generated by other tools from Google, OpenAI, and Adobe.

  • As it’s harder to detect AI-generated video or audio, they’ll ask users to self-label this content, threatening penalties (including content removal) if they fail to do so.

  • This comes after Meta’s independent oversight board slammed its current media policy, calling it "incoherent and inappropriately focused on how content has been created."

🤨 Why you should care: As the lines between real and fake AI-generated media become more blurred, Meta is taking these steps in a bid to tackle AI fakery and galvanize the tech industry, encouraging others to follow suit.

Do you think Meta's labeling strategy will stop image fakes?

Login or Subscribe to participate in polls.

 

In Partnership with Brilliant.

Brilliant's bite-sized, interactive learning makes it easy to level up in minutes daily—whenever, wherever.

Brilliant is your shortcut to:

  • Thousands of quick, visual, hands-on lessons in AI, data science, programming, and more

  • Interactive problem-solving is a more effective (and incredibly fun) way to learn

  • Bite-sized learning makes it easy to build knowledge you can put to work

Forget spending hours watching videos or dropping $$$ on classes. Join over 10 million people and start your 30-day trial today. Even better, AI Tool Report readers can get 20% off an annual premium subscription.

ROBOCALLS

  • Investigations began after voters in New Hampshire were urged to save their ‘votes for the November election’ by a fake robocall (made by ElevenLabs) imitating Biden in January.

  • The fake calls were traced to telecoms provider Lingo (which has been conducting illegal call operations for years), who transmitted the calls to New Hampshire residents.

  • But it was Life Corporation—owned by Walter Monk, who delivered illegal pre-recorded ads to residential phone lines in 2003—that commissioned the calls.

  • Life Corporation and Lingo have been issued with ‘cease and desist’ documents (warning them of potential law enforcement) and they could be charged with fraud.

✅ Designed to work alongside our new job board—“Work in AI” (launching on Monday 12th February)—our Work In AI GPT will help with every aspect of job applications, from perfecting resumes to mastering interviews.

Edition 2 of 3: The Practice Interview

When To Use: Because you optimized your resume (using Work In AI GPT ‘Optimizing Your Resume’, you’ve now been invited for an interview.

  1. Step 1: Start A Practice Interview: Click on option 1: "Interactive Practice Interview." You will be asked to provide more information about the job you are applying for and to upload your resume, to create the perfect practice interview.

  2. Step 2: Answer the 10 Job-Specific Interview Questions: Sample question: Can you describe a project where you applied your engineering skills to solve a complex problem? What was the challenge, and how did you approach it?

  3. Step 3: Receive Your Score And Feedback: After answering your questions you will directly receive feedback on how you did and what you could improve.

By practicing your interview with the Work In AI GPT you are going to be extremely well prepared, and the actual interview will feel like a walk in the park.

P.S. Get ready for our next edition, where Work in AI will craft your perfect Cover Letter, tailored to any role, using your resume and job vacancy description.

HEALTHCARE

  • OpenAI is leading a $70M Series B funding round (with input from A16z and Optum Ventures) to help Ambience Healthcare develop its AI healthcare platform.

  • Their platform aims to streamline the administrative workload for clinicians and address medical professionals' complex data processing needs.

  • It will offer tools to analyze conversations and documentation, generate automatic referrals, complete after-visit patient summaries, and assist with appointment preparation.

  • This funding round takes Ambience Healthcare’s total funding to $100M, allowing them to build on their previous venture (Remedy Health) which leverages AI for clinical diagnosis.

SAFETY PROTOCOLS

  • Global AI companies (including Google, OpenAI, and Meta) have clashed with the UK over AI safety protocols, urging the country to speed up the implementation of AI safety tests.

  • After making a voluntary commitment in November to change their AI models if safety flaws were found during testing, they’re now seeking clarification of the results.

  • The government has established that AI safety testing is underway, but that companies should take accountability for testing themselves, before releasing their AI models.

RETAIL

  • IKEA has introduced a GPT-powered shopping assistant that leverages AI to offer personalized furniture and décor visualizations, based on prompts.

  • The AI shopping assistant considers room dimensions, personal style, and sustainability preferences before offering recommendations and access to IKEA products.

  • This new feature is currently only available to ChatGPT Plus users in the US, but IKEA plans to roll out access to other countries, over the coming months.

Over the past year, AI Tool Report has worked with Hubspot, Brave Browser, SurfShark, Guidde, Graft & many more—presenting their AI products/services to over 490,000 AI enthusiasts, boasting:

  • A consistent 52% open rate

  • Best-in-class ad performance

  • A growing ecosystem of AI products/services

We read your emails, comments, and poll replies daily.

Hit reply and let us know what you want more of!

Until next time, Martin & Liam.

What did you think of this edition?

Login or Subscribe to participate in polls.

Subscribe to keep reading

This content is free, but you must be subscribed to AI Tool Report to continue reading.

Already a subscriber?Sign In.Not now