Generative adversarial networks, or GAN, are sprouting up all around us. From the cover of Dazed Beauty’s Issue 0 (Kyle Jenner’s makeup was created by GAN), and a website that generates infinite images of people who don’t exist, to ThisStartupDoesNotExist.com, which is a website that creates, at every click of the refresh button, a new landing page for an imaginary startup here to revolutionise your… something.
GAN works by running two classes of neural networks into one final output—meaning that it takes real data and feeds an algorithm that distorts it in a formulaic way in order to create more variations that are anchored in real traits, but are completely fake. In many ways, GAN can be used for positive developments such as advancing virtual realities in video games or even aid in the visualisation of design across fashion to interiors as well as urban planning.
In response to the growing use of GAN in our day-to-day, a movement of websites has risen, examining the many aspects of our life that GAN is capable of affecting. Using a variation of the URL address This XX Does Not Exist, ThisStartupDoesNotExist.com is the latest addition to this movement (alongside ThisWaifuDoesNotExist.net and ThisPersonDoesNotExist.com). The site runs GAN to generate endless fake startups websites that look and sound real, but are in fact created in a matter of seconds.
What makes ThisStartupDoesNotExist.com so intriguing is its accurate depiction of a startup movement that has seen companies founded by the thousands. All the companies somehow follow a similar formula which was most likely subscribed to them from accelerators and advisors in the chase for VC investment. This site accurately exposes this formula in its ability to use GAN to generate endless startups that seem, on the surface, legitimate.
The layout resembles that of a classic landing page of a 2019 startup. A large hero image with the startup name and tagline greets the users as they are first introduced to the company. “Econdr, Build hardware that unlocks new use cases”, is one startup’s name and mission. “Finise, Streamline a process through tech”, reads another. The names of the fake companies are well in line within the trend of startups using one word names that oftentimes take on a go-to suffix like -ly or -ify (Contently, Spotify), drop a few vowels (Grindr, Flickr) or adapt to a magician’s lexicon like in the case of (Shazam, Hulu, Venmo).
After users are introduced to the company through its landing page, name and tagline, they need only scroll down to discover the team section. For any aspiring startup, this is a crucial part of the business, one that Venture Capital companies examine closely as they chose who to invest their money into. A winning team of CEOs, COOs, CFOs and CTOs is key. While it isn’t clear exactly where ThisStartupDoesNotExist.com pulls the images for its fake team members, it is assumed that these are taken from ThisPersonDoesNotExist.com and attributed a random combination of names.
“Finise is more than just Streamliner. Community, Native and Served are just a few characteristics of Finise. But we offer a lot more!” reads the tagline above the prices and planning further down the page. Following which users are presented with several price plans for various uses—from personal to enterprise. And finally, the whole faux startup is sealed with a stamp of approval using quotes from fake clients currently using the service, a newsletter sign up option, and a ‘get in touch’ form to fill.
This website of endless fake startups is a subtle and humorous yet powerful way to highlight the absurdity present in the startup era. According to a 2015 report by Forbes, approximately 90 percent of startups fail, and it feels as though ThisStartupDoesNotExist.com is bleak criticism of the culture behind an industry of ‘disruptors’ who rely so heavily on VC funding and thus conform to an often destructive ethos of growth, inflated markets and disruption where it is not always necessary. Surely GAN’s ability to generate endless fake startups in a matter of seconds should be seen as a wake-up call that we should take this industry with a pinch of salt.
2018 has seen a disturbing surge in the effects of climate change, which wreak havoc across every continent. Whether it’s due to unbearable droughts or perilous floods, entire communities around the world undergo the painful process of bouncing back from natural disasters and readjusting their lives to the rapidly changing climate and topographical conditions of the Earth. Although human activity and global industrialisation are to blame for the destruction of our planet’s ecosystem, it seems that, at least to an extent, a solution to the problem could emerge from technological advancements. Numerous initiatives are currently underway to utilise Artificial Intelligence (AI) and big data in order to both prepare for the changing landscape of coastal areas and increase the efficiency of human activity in order to minimise its adverse impact on the environment.
One Concern is a start up managed by Craig Fugate, the former head of the Federal Emergency Management Agency (FEMA). The company employs AI and machine learning technologies to provide local authorities, community leaders, and decision makers with highly precise data regarding upcoming natural disasters, which will allow them to adequately prepare for their impact. One Concern supplements their high tech services with an in-person training to help their clients assess, respond to, and recover from natural disasters. In an interview for Scientific American, Fugate states, “If we can make the risks more definitive, we can at least start to get more control over our destiny—versus it being inflicted upon us each time a storm hits.” Fugate further argues that as part of its mission, One Concern advocates for an equal and fair transformation of coastal areas which over the next decades will have to plan for massive relocation of its population. Fugate warns against a scenario in which the burden of moving will be placed solely on lower income communities, and that big data and AI will be exploited by biased entities and developers seeking to maximise profits. The start-up considers all these factors when providing assistance and consultation to its clients.
Disaster prevention is only one aspect of the High Tech vs. Climate Change trend. AI For Earth is an initiative developed by Microsoft seeking to employ AI in an eco-friendly manner. The programme is aimed at providing individuals and organisations with technology that allows them to utilise climate data in order to increase the efficiency of their water and energy usage. Specifically, the programme uses AI in order to translate and process raw data of environmental systems so that farmers and agricultural companies could divide it into categories that make sense to them. In collaboration with the National Geographic Society, AI For Earth will be giving grants to a selected group of AI app designers and projects they believe could help perfect their product and furnish members of the agricultural sector, scientists, and sustainable initiatives with valuable and usable data regarding climate patterns.
Despite these tech advancements, here still lies an inherent risk in profiting off of such critical data compiled by AI and machine learning, as overtime such initiatives could be primarily concerned with monetary gains as opposed to tackling climate change and providing equal assistance to all those who are vulnerable to its effects. And so, as such technology is developed, it is crucial that both national and international organisations and bodies oversee its usage and ensure that its application benefits more than a select-wealthy-few.