On Friday a flawed update by endpoint virus detection and response provider CrowdStrike triggered a global outage that crippled the travel industry’s IT infrastructure. Airports ground to a halt, airlines canceled flights en masse and frustrated travelers faced long queues and delays.
The financial repercussions will be substantial. The damage to customer trust and the industry’s reputation will be even more profound. The incident is a reminder of the golden rule-of-thumb that everything in technology can break – and eventually will break.
There are segments of the industry that have fully embraced this principle by avoiding single points of failure. Any modern commercial aircraft, for example, will have a cascade of redundant backup systems (usually three). You can fly and land an aircraft with only one engine and bring it to a standstill on a runway just by using the wheel brakes.
For example, the “Master Minimum Equipment List” for an airworthy Airbus A380-800, as published by the United States Federal Aviation Association, is 216 pages long! We live in an era where fail-safes for fail-safes for fail safes are mandated.
And yet, many still don’t possess this mindset. Now, the future of travel hinges not just on recovering from this massive IT failure but also on what comes next.
Besides having to re-deploy staff to manual IT operations, travel companies will have to rethink their system setup to prevent such a catastrophic failure from happening again anytime soon.
Embedded artificial intelligence systems can play a proactive role in building resilience and ensuring seamless operations becomes the norm, because the next black swan technology event is always just over the horizon.
We must be better prepared.
Today’s AI-powered solution: RAG-based script deployment
Something companies could implement tomorrow is a RAG (retrieve, augment and generate) AI framework. This system is designed to remotely diagnose and resolve CrowdStrike-like issues, utilizing several AI technologies working in concert.
A RAG framework collects data (including logs, system metrics etc.) from endpoints in the systems it monitors. It then uses AI-driven tech like BERT [bidirectional encoder representations from transformers] and GPT to add context by analyzing the data.
BERT and GPT then scrutinize this enriched data to detect anomalies and identify potential issues. Advanced language models like GPT4o, Claude, or Llama then generate human-readable insights and recommendations, while a library of custom scripts executes predefined troubleshooting steps remotely.
Finally, the RAG framework’s strict security protocols and continuous learning mechanisms ensure compliance and data protection.
This integrated approach could have been able to prevent or significantly mitigate the effects of the CrowdStrike outage.
In the future: the promise of embedded AI
Embedded AI, also known as Edge AI, has gained some momentum this year. Just a few weeks ago Apple announced the launch of the fully embedded “Apple Intelligence” on its devices and Meta is already running Llama 3 as part of its Instagram and WhatsApp apps.
By integrating AI directly into systems and devices (versus running it solely on the cloud), travel companies can achieve unparalleled levels of system independence, predictive maintenance, autonomous response and continuous learning.
All of the above can be used to build more resilient systems. Black swans can’t be stopped, but their impacts can be heavily mitigated as a result of this process.
Here are some of the core benefits:
- Embedded Predictive Maintenance: AI can continuously monitor system performance, predict potential failures, and trigger preventive measures before issues escalate (or cascade, as they did in recent days).
For example – a computer mainboard is running hot. The AI script gets notified by the hardware sensors that the temperature is off and starts checking the load of its top running programs. It then identifies that there is an element of the code that is stuck and has a memory leak using a nested AI logical search.
It checks via API-connect for known issues and what needs to be done to apply a fix or identifies the latest working backup date and escalates to an end-user/admin for action. Or, if it doesn’t receive a response at all, it can define its own action within a certain timeframe.
- Device-level autonomous response: In the event of disruption embedded AI can take immediate action by autonomously initiating recovery protocols. This reduces downtime and keeps critical systems running, minimizing the impact on operations and customers.
- Single-node continuous learning: Embedded AI systems can learn from each incident, constantly improving their ability to detect and mitigate future threats. Each node of the system could then develop further measures and share them. This dynamic learning capability means that the more the nodes are exposed to issues, the better the overall AI foundation becomes at preventing similar problems.
- Stand-alone enhanced security: With real-time threat detection and mitigation, AI can bolster the security of travel IT systems. Each node becomes better able to protect itself against cyber threats that could otherwise exploit system-wide vulnerabilities during outages.
- Dynamic risk assessments: AI can provide continuous risk assessments, identifying vulnerabilities as they emerge and suggesting proactive measures to address them. A network of AI-embedded nodes can work with a “master” AI to bolster its response to possible incidents.
- Employee empowerment: Training employees to work alongside AI tools ensures that human expertise complements technological advancements, creating a robust defense against a Black swan event.
Looking ahead: The future of travel in an AI-enhanced world
The global CrowdStrike outage is a wake-up call – a challenge which cannot be ignored. Using embedded AI will allow the travel industry to move beyond mere recovery and build a future where disruptions are anticipated and mitigated.
We can’t prevent black swans. However, we can significantly reduce the risk of their re-occurrence and the damage they cause.
The future of travel lies in intelligent, proactive, and resilient systems powered by AI. The question should no longer be “if” the industry will adopt these technologies, rather “when,” and how quickly it can do so to safeguard its operations and earn back the trust of customers.
This article originally appeared on PhocusWire.
