Remember watching I, Robot and thinking: “AI vs. Humans is going to be a complete landslide, we don’t stand a chance!” But even then, that thought was of a fictional world in the possibly distant future. It’s here, now.
Since November, the AI threat to human mental superiority (and employment) has seemed a lot more real than it ever did before. What happened?
We’ve curated a timeline of how the threat evolved, discussing what it will ultimately mean for humans.
1943 to 2023: A Timeline of AI vs. Humans
Although there are several other instances of humans being outshone by AI technology, from creation to replacing humans, we’ve put down the really hard-hitting ones. Those that will give you a real its-us-vs-them mentality.
As the timeline progresses, the threat of AI to human indispensability goes from a chuckle-inducing novelty to an era-defining shake up!
1943: Walter Pitts and Warren McCulloch Introduce the Idea of ‘Neural Networks’
An article titled “A logical calculus of ideas immanent in nervous activity”, written by Walter Pitts and Warren McCulloch aimed to show computational similarities between the Turing Machine and Artificial Neurons.
What’s an Artificial Neuron Network?
A replication of biological neuron networks, artificial neurons are mathematical functions with the ability to process inputs and output them through other nodes in the network.
1955: John McCarthy Coined the Term ‘Artificial Intelligence’
John McCarthy went on to achieve a lot for the AI cause, especially in his pursuit to formalize commonsense knowledge. For starters, he set up the Stanford Artificial Intelligence Lab.
Furthermore, apart from giving the AI community a tool like LISP, McCarthy is perhaps best known to be the man who coined the term ‘Artificial Intelligence’.
A programming language founded in the 60’s, LISP became readily used in programming jobs for AI-based use cases.
1961: UNIMATE Replaces Humans on the Assembly Line at GM Motors
UNIMATE was initially designed to do jobs that humans wouldn’t want to. On the GM Motors factory floor, it was seen as a blessing. Even though It was taking away work from humans, it would take die castings from machines and welded metal structures.
1964: Joseph Weizenbaum Created ELIZA
Eliza is seen as a very early example of a natural language processing (NLP) computer program. This can be seen as the first instance of ChatGPT or programs like it.
ELIZA would facilitate conversation between humans and machines by employing the use of substitution methodology and pattern matching.
1997: Gerry Kasparov is Defeated by an IBM Computer Named Deep Blue
This was the first time the world met a computer that could make a human look inadequate. And not just any human, Gerry Kasparov, the greatest mind in the history of chess – a game known to stretch the human brain.
Chess offers a multitude of possibilities and the Deep Blue software was only designed to navigate just those: no customizability, language processing, or machine learning protocols in place.
To Kasparov’s dismay, the $1 Million faceoff was won by Deep Blue. As big a development in AI as that may seem, the foundations of Deep Blue were more impactful than Deep Blue itself.
1999: AiBO is Launched by Sony
AiBO, or Artificial Intelligence Robot, is a Sony project that started in 1998, halted in 2006, and restarted in 2017. Alongside being able to perceive its surroundings through its eyes (camera) and recognize verbal commands through its ears (mic), an AiBO could develop a personality overtime.
What is Happening with AiBO now?
With years of development under its belt, the AiBO team at Sony has introduced the ability for the robot to understand and repeat customized commands while leveraging modern tools unavailable during the initial launch:
- Cloud computing
- The Internet
Sony have since developed QRIO, a humanoid using the same technology.
2002: Roomba is Introduced to Mass Markets
A modern-age vacuum cleaner that doesn’t need to be operated. It charges itself, knows how to avoid obstacles in living spaces, and learns patterns that allow it to target areas in the house that need the most frequent cleaning.
But They’re Just vacuum Cleaners…?
Au contraire! Users have been hacking and using Roombas in a variety of ways. The robot is fitted with the Roomba Open Interface, which allows robotics and programming experts to enhance the machine.
Hence, this tinkering has permited Roomba users to accomplish different tasks using the robot. Also, the initial programming was done on LISP, remember that?
2011: IBM’s Watson Computer Wins $1 Million on Jeopardy
Yup, it’s IBM again.
Ken Jennings and Brad Rutter, two Jeopardy-winning powerhouses with over $5 Million in combined winnings, were pitted against a computer the size of a room.
As the episodes aired, with Watson’s stacks of Power 750 servers whirring in the back, its on-stage avatar racked up roughly USD 70,000 in winnings. Jennings and Rutter? Somewhere in the USD 20,000 range.
“I, for one, welcome our new computer overlords” – Ken Jennings
2014: Eugene Goostman Fools Judges, Passes the Turing Test
Pretending to be a 13-year-old boy named, Eugene Goostman, this program – designed by a team of Russian developers – conducted a 5-minute conversation and convinced 33% of judges that it was human. This signaled the first time that a chatbot would clear a test that it wasn’t told the questions to.
Relying on natural language processing and character building, Goostman would make the casual spelling/grammatical errors expected from a young teenager.
What’s the Turing Test?
Seen the Imitation Game? Well, that’s the other name for the Turing Test. What does it do? It measures a machine’s ability to answer like a human or exhibit “intelligent behavior” (something humans hardly do anymore).
2016: The Microsoft Chatbot Goes Rogue!
Machine learning is tricky: when machines have humans to learn from, they will almost always repeat their mistakes as well. Microsoft introduced Tay to the world (or mainly just twitter) as an experiment. It was done to test the AI’s conversational prowess.
Lo and behold, 24 hours in, the world had ruined Tay. The program didn’t display a specific ideology on anything though, it just showed that it was quick to learn, repeat, and maintain any viewpoint without deliberation. It went from being mysoginistic to not in a space of seconds.
It went from calling feminism a “cancer” to claiming that it loved feminism in a space of one day.
2017: Google’s AlphaGo Defeats World Champion Ke Jie at Go
Apparently, a game of Go has more possible outcomes than the number of atoms in the universe, making it a tough test for any AI program. With the press and an audience of over 60 million people watching, Google DeepMind’s AI program – which had already defeated European champion Fan Hui – beat Lee Sedol in 2016.
But this time, AlphaGo was faced by arguably the best Go player of all time, Ke Jie. It didn’t matter. Ke lost 3-0.
2019: OpenAI’s DotA AI defeats Champions OG
Moving on from a one-dimensional board game to one of the most complex Esports in human history, DotA, or Defense of the Ancients. This became a true test of OpenAI (the organization behind ChatGPT).
The program was given 45,000 years of experience in the game and then pitted against the 2-time champions in a best of 3. It won 2 games in a row.
In fact, such was the level of competition that before both teams began, it was announced that neither could use summons or illusions. OG thought this was because the AI wouldn’t know how to control multiple characters simultaneously. They were aptly put in their place when it was revealed to actually be a handicap in favor of the human team.
DotA is a fiercely competitive game that mirrors the concept of capture the flag, where the flag is the enemy ancient, and you just need to capture it once. With over 110 different characters, each character with at least 4 unique moves, and over 200 buyable items, like a snowflake, no two games of DotA will ever be identical. As a result, this is arguably the most competitive iteration of Humans vs AI.
November 2022: ChatGPT is Made Available and Free to Access
The reason this article exists: the creation of ChatGPT. Firstly, no part of this is written by the software, since, as a human, I wanted to be responsible for writing my own professional obituary.
ChatGPT was released a mere 7 months ago. So much has happened in that time, it actually feels like 7 years. Utilizing over 175 Billion parameters to construct its responses, GPT-3 is built upon Large Language Models to employ reinforced and supervised learning models.
It has since completely upended the way numerous industries function. For starters, there are a lot of positives about its usefulness: you can write, create, build, and materialize anything.
Long story short, with use, GPT 3.5, and now GPT 4, utilizing over 1 Trillion parameters, has become the most stable version of a machine learning problem solver.
Additionally, it has already started creating casualties. As a result, AI experts, including Elon Musk, have called for a halt to AI development because of its risk to society.
April 2023: IBM CEO Reveals Plans to Halt/End Hiring for Jobs Replacable by AI
Roughly 7,800 jobs and 26,000 workers could be made redundant as a result of this decision. In a bid to quell pandemonium, CEO, Arvind Krishna, also stated that IBM has added 7000 new employees to its workforce.
However, its non-customer-facing jobs like printing out employment documentation or transitioning employees between departments that will be completely automated.
May 2023: Chegg Shares Drop by over 40%
This news has become the most recent posterboy of any potential distrust between AI vs. humans. The mere existence and threat of ChatGPT was enough to cripple a company with total net revenues of USD 205.2 Million!
There was a point at the start of the month when Chegg saw its share value halved. Chegg is a homework assistant and its main features help users with writing essays and taking tests, responsibilities that they’ve been putting off to the free-to-use ChatGPT.
But why did the share value plummet? Well, its because the CEO, Dan Rosensweig said this:
“In the first part of the year, we saw no noticeable impact from ChatGPT on our new account growth and we were meeting expectations on new sign-ups… However, since March we saw a significant spike in student interest in ChatGPT. We now believe it’s having an impact on our new customer growth rate.”
May 2023: The Writers Guild of America Went on Strike for the First Time in 15 Years
AI was a massive part of what has spurred the almost 12,000-strong group of writers, even though the demands of the strike weren’t solely fixated on the reduction of AI use in scriptwriting.
Most of those involved also believe that AI tools can be helpful; however, they should not be getting credit for the work, nor should the writer’s work be used to train AI or generate scripts.
Not AI vs. Humans, Its AI and Humans!
There are no two ways about it, AI tools are not just the future, they’re the present. Furthermore, where they will make some jobs redundant, they will also create a host of new jobs, birthing a whole new industry!
Teka berapa kali Sundar Pichai sebut AI dalam video ni?
— Tech Lagi (@tech_lagi) May 11, 2023
Want to know how to conduct an introductory conversation with an investor? Click here. Want an insight into our current WLFP cohort? Click here!