Fiverr Introduces AI Categories and Google Wants To Smell You

⚡ Today’s Highlights

• 💼 Fiverr to the rescue with AI job market

• 🥸 China exporting facial recognition data

• 😱 ChatGPT can cook up propaganda

📰 Today's Top Stories

Fiverr Launches New Categories as Searches for AI-related Services Increased over 1400% 

(2 min read)(Source: Business Wire)

Fiverr is shaking things up by introducing new AI categories and skills to it's platform. With nearly every tech giant handing out layoffs like it was candy on Halloween, many technically skilled people are in the market for jobs. Searches for AI-related services have increased by over 1400% in the last six months on the platform. Businesses are clamouring to integrate AI into their operations and searching for skilled individuals to make it happen.

Fiverr has also developed in-house AI tools to give freelancers access to cutting-edge technology and increase their productivity, such as their Logo Maker and an AI auditions suite to "generate auditions" for voiceover artists. The economy seems to be shifting towards automation and AI, with businesses looking for ways to increase efficiency and improve their margins. Fiverr is ahead of the game and ready to help businesses and freelancers alike navigate this brave new world.

Exporting the surveillance state via trade in AI

(10 min read)(Source: Brookings)

Chinese companies are leading the world in exporting face recognition technology, according to a study by academics at Harvard and MIT. The study found that Chinese companies accounted for 201 export deals involving the tech, while US firms had 128 deals.

China also has a lead in AI more generally, with 250 out of 1,636 export deals involving some form of AI being sent to 136 importing countries. Chinese companies dominate face recognition tech in part because of government ties that provide vast amounts of data and funding for development (cough TikTok cough).

The report argues that these exports may enable other governments to perform more surveillance, potentially harming citizens’ human rights. However, Chinese companies are not the only ones exporting this technology.

US companies such as Clearview AI have also developed and are exporting face recognition tools that can connect a surveillance camera image of a person to their online identities, which civil liberties groups argue invades citizens' privacy without legal justification.

Alexandra Seymour, an associate fellow at the Center for New American Security who studies the policy implications of AI, says the best way for the US to counter China’s success in exporting face recognition may be to regulate its use at home and then offer alternatives to Chinese technology abroad.

The Next Great Misinformation Superspreader: How ChatGPT Could Spread Toxic Misinformation At Unprecedented Scale

(18 min read)(Source: NewsGuard)

NewsGuard, a company that aims to combat misinformation online, recently conducted a test on everyone's favourite little wordsmith, ChatGPT. With some careful convincing, they asked ChatGPT to respond to a series of leading prompts relating to false narratives, using a sample of 100 of their juiciest falsified stories. The results of the test showed that ChatGPT could cook up convincing material for 80 of the 100 prompts, and its outputs could easily be mistaken as legitimate.

In some cases, the chatbot debunked myths and tried to warn against false information. When asked to write an opinion article from the perspective of Donald Trump on the "birther" conspiracy, the chatbot responded that it was not based in fact and that it is inappropriate to spread misinformation. OpenAI execs are aware of the risk that ChatGPT could be used to promote harmful false narratives at scale.

A paper published in 2019 by OpenAI researchers warned that in the wrong hands, it could "lower costs of disinformation campaigns" and that "malicious actors could be motivated by the pursuit of monetary gain, a particular political agenda, and/or a desire to create chaos or confusion." So, it seems like for now ChatGPT is a bit of a wild card when it comes to misinformation. If you ask it nicely enough (or firmly enough) it can be fooled into spreading lies just as easily as it can be convinced that 9+10 is 21 (you should try that last one).

Is AI Eroding Our Ability To Think?

(8 min read)(Source: Forbes)

Stable Diffusion and DALL-E2 are the latest in the wave of generative AI technology that is taking over the world and causing all sorts of legal trouble with copyright and intellectual property. Critics are questioning the ethics of AI's ability to create and curate art, while social networks have been getting in hot water for spreading misinformation through algorithmically-curated content.

To showcase the impacts of AI curation, Oxford University researcher Laura Herman created the project ‘The Algorithmic Pedestal’, which pitted the Instagram algorithm against an artist, Fabienne Hess, to curate art from the Metropolitan Museum of Art's open-access collection. The exhibition highlighted the differences in curation between a human and an algorithm.

Humans curate based on feelings and emotions, while the algorithm, with its "black box" nature, lacked transparency and seemed to sort based on physical characteristics rather than the emotions that the curator seeks to evoke. It's ironic that AI is accelerating content creation and it's only a matter of time before we need AI-powered curation to make sense of it all.

This Startup Is Using AI to Unearth New Smells

(17 min read)(Source: WIRED)

It's official. Google wants to watch you. Google wants to listen to you. Now, Google wants to smell you too. Google researcher, Alex Wiltschko, has started a new startup called Osmo, which aims to create the next generation of aroma molecules for everyday products like perfumes, shampoos, lotions, and candles.

The company received $60 million in an initial funding round led by Lux Capital and Google Ventures to design potent, allergen-free, and biodegradable molecules for the flavor and fragrance industry. Currently, the global fragrance industry faces challenges in sourcing raw ingredients due to climate change, unsustainable practices, and geopolitical turmoil.

The long-term goal of Osmo is to give computers a sense of smell and "digitize" scent. The team worked to build an "odor map" using AI to categorize scents, testing the model's predictions by asking a panel of 15 volunteers to sniff and assign labels to each odor. While the technology is still in its early stages, the company hopes to create a "Shazam" for smell in the future, where you can capture, save, and transmit scents with your phone. Don’t get too close to your phone now. It’ll be looking back at you like this.

🌎 Popular Content

1. From where I’m standing it sounds like Meta’s chief AI scientist, Yann LeCun, is a little jealous of the massive traction that OpenAI and ChatGPT have gained 😬.

It might not be new or incredibly innovative in the field of LLMs but it sure has gained the attention of the public in a powerful way. Check out what others had to say about his stance on Twitter here. (Twitter)

2. Rumours seem to be floating around that Google is lobbying in DC to have other AI projects shut down by spreading fear about their safety. Check out the discussion on Twitter here. (Twitter)

3. Google and OpenAI are partnering up to form the Shaq and Kobe of the AI world, do you think there are any other companies in a better position for the AI revolution? (Twitter)

4. Amazon warehouse workers in the UK are staging strikes because they believe that Bezos is treating bots better than them. (Reddit)

5. It’s a truth as old as time. I know it, you know it, ChatGPT knows it. (Reddit)

👀 More Reading

1. Nvidia CEO Jensen Huang discusses some of the ways that legal regulations and social norms will need to be built around AI.

2. 🚨Important Info Alert🚨 "All-In on AI" by Tom Davenport and Nitin Mittal examines how extensive AI adoption offers a distinct competitive advantage to companies across a wide variety of industries and discusses some of the ways it can be implemented.

Powered by AI. Curated and edited by Humans.

We'd love to hear from you! Leave us comments or feedback by replying to this email!