Artificial Intelligence (AI) describes technologies that can make informed, non-random decisions algorithmically. It has many current and potential applications, it is the current pinnacle of humanities ceaseless drive towards greater and greater efficiency. In particular regard to OSINT though, it enables humans to collect, analyze and interpret huge sets of data, data sets so large that it would be entirely unfathomable to even approach them without machine assistance.
Everyone knows AI is shaping their world in one way or another. But often the changes are subtle, gradual and go unnoticed. Very few of us know what actually goes on behind the steel doors of the big tech companies like Alphabet, Facebook, and Apple. And yet we interact with their AI systems on a daily basis and those systems have huge power over our lives. In this article, we take a look at some of the key ways AI is being used today and how it will become increasingly important as our technologies improve.
5 Ways AI is Shaping the World
1. Improving and optimising business processes
The very first robots in the workplace were all about automating simple manual tasks. This is the age of factories and production lines. Today though, it’s not manual tasks that robots are taking over. Instead, software-based robots are taking on repetitive tasks carried out on computers.
Initially, this was limited to automating simple repetitive tasks, such as “send follow up email 2 if no response after 3 days”. This has already reduced admin tasks and improved business operational efficiencies immeasurably. The next step though is the use of AI technologies to further alleviate some of the more labour intensive ‘intelligent’ tasks such as data gathering, aggregating and analysis, leaving people to spend more time on complex, strategic, creative and interpersonal tasks.
2. More personalization will take place in real-time
Big tech companies are already using data to personalization services. Google Discover, for example, is a feed based on a complex algorithm which reads your online history and tailors the news feed to your particular interests. Other big tech examples are Spotify and Netflix which use AI to suggest relevant media based on your historical behaviour.
This technology is constantly being evolved and is probably one of the most noticeable in our day to day lives. The end goal is a system which can almost perfectly predict your desires and needs, an outcome none of us are likely to protest against. On the other side of the same coin though is the use of that very same data to target individuals with hyper-relevant ads. This practice can often seem intrusive and is one of the driving forces behind the adoption of VPN’s.
3. AI in the creative space
Some things are still, even in 2020, better handled by humans. That being said AI technologies are now beginning to encroach on the creative spaces. Scorsese's, The Irishman, is one example of this, where Robert De Niro was de-aged on-screen using AI technology.
There are additional uses though, for example, AI is being used to edit video clips for the purposes of spreading misinformation, and often these edits are incredibly hard to spot. This has led to a new sector of cybersecurity which requires AI technology to spot AI-generated or edited video and audio files.
4. Increasing AI in Cybersecurity
Even as data grows and is used to progress the development of AI this simultaneously opens up new avenues for exploits by threat actors. For example, AI can be used to create and automate targeted ‘intelligent’ phishing campaigns. AI-supported cyberattacks though have the potential to go much further. As such, increasingly advanced AI is needed to combat the evolving cyber threat landscape.
Related: How Machine Learning is Changing Modern Security Intelligence
5. AI learning to perfectly emulate humans
Anyone that keeps their eye on the work that Google is doing will know about their 2019 update, BERT. A natural language processing (NLP) framework which is designed to better understand context and intertextual reference so that they can correctly identify both the searcher's intent as well as the intent behind any content created.
One of the key challenges that faces AI right now is idiomatic or referential speech; language that has more depth of meaning, for example, determining the importance of the concept of a mother, or understanding a phrase like “six feet under”. Our current research and development project at Signal is one example of the practical applications of overcoming this challenge. It involves using machine learning to enable our software to understand the intent behind text, even when ‘hidden’ behind challenging language like idioms, to more accurately identify threats.
As these natural language processes advance, so too will conversational AI bots, to the point where, because of the range and complexities of their answers, you would be forgiven for mistaking them as human.
The Future of AI and what that means for OSINT
Artificial Intelligence, machine learning, and automation have already revolutionized intelligence gathering. With OSINT tools like Signal security teams and intelligence agents can effectively and efficiently monitor the open, deep, and dark web, setting up customized alerts based on searches that leverage boolean logic. Machine learning takes this intelligence to the next level. It allows for vast amounts of data to be collected, aggregated, and for all the irrelevant hits to be essentially culled, supplying the security team at the end with actionable, relevant intelligence.
Humans play an essential role in this new intelligence lifecycle. In defining the search terms to match security strategies, analysing the end date the system feeds back, reassessing the searches based on the new evidential data and implementing appropriate responses. This is a key role that will no doubt evolve as the technology becomes more accurate, reducing inefficiencies in process.