Jump to content

Draft:LXT AI

From Wikipedia, the free encyclopedia
  • Comment: If you're hoping to rely on the items listed in 'References' to establish notability, it would be very much preferred that you cite them. DoubleGrazing (talk) 12:25, 8 May 2025 (UTC)
  • Comment: Even if this was rewritten, it does not meet the notability criteria. As far as the Forbes sources, see WP:FORBESCON. They should not be used nor most of the other sources. S0091 (talk) 20:51, 5 May 2025 (UTC)

LXT
Company typePrivately held company
IndustryArtificial intelligence, Data services
Founded2010; 15 years ago (2010)
FoundersMohammad Omar (CEO)
HeadquartersToronto, Canada
Number of employees
350+ (2025)
Subsidiariesclickworker (since 2025)

LXT is a Canadian artificial intelligence (AI) data services company that provides multilingual and multimodal data to support the training and evaluation of AI systems. Headquartered in Toronto, the company works with organizations across industries to supply data for applications such as natural language processing (NLP), computer vision, and generative AI.

History

[edit]

LXT was founded in 2010 and initially focused on language data for speech recognition technologies.[1] Over time, it expanded into image, video, and multimodal data annotation. In December 2024, LXT acquired German crowdsourcing company clickworker, which operates a global contributor platform.[2]

Research

[edit]

LXT has published annual executive surveys on AI maturity among enterprises. Its 2023 and 2024 editions, titled "The Path to AI Maturity", provided insights into trends in AI investment, governance, and adoption across industries.[3]

Services

[edit]

LXT provides data services used in the development and evaluation of artificial intelligence systems. Its work focuses on collecting, annotating, and validating data across modalities including text, audio, image, and video. According to the company, it provides coverage for over 1,000 language locales globally and has experience working in more than 145 countries.

LXT’s services include multilingual data collection, speech transcription, image annotation, and prompt evaluation for large language models. In addition to training datasets, the company also offers model evaluation and benchmark testing. These services are typically delivered under a managed services model, where data quality and workflow security are managed internally.[4]

Clients of LXT include global technology companies developing AI systems for applications such as voice assistants, computer vision, autonomous vehicles, and content moderation.[1]

Importance of AI Training Data

[edit]

Supervised machine learning systems rely on large volumes of annotated data to learn patterns, make predictions, and generate outputs. The quality, diversity, and quantity of this training data significantly influence a model’s performance, generalizability, and fairness.

Training data is required across a wide range of domains, such as language translation, image classification, speech recognition, and text generation. To reduce bias and ensure applicability across different populations and contexts, datasets must often represent a variety of languages, dialects, visual environments, and social backgrounds.

To meet the increasing demand for such large and varied datasets, organizations commonly rely on crowdsourcing methods to collect and annotate training data. This approach leverages distributed contributors from around the world to perform tasks such as data collection, labeling, transcription, and evaluation.[4] Crowdsourcing can be particularly effective for gathering multilingual or culturally specific data at scale.[5]

In many data pipelines, crowdsourced annotation is integrated into a broader human-in-the-loop (HITL) framework. In such systems, human input is used to guide, correct, or validate the outputs of machine learning models during training and deployment. This method is particularly relevant for tasks that involve ambiguity, subjective judgment, or ethical considerations, such as sentiment classification, content moderation, or dialog evaluation.

Global Crowd

[edit]

Following the acquisition of clickworker, LXT expanded its operational reach by integrating a global contributor base. The combined entity reportedly manages a workforce of over seven million contributors in more than 200 countries and regions.[2]

See also

[edit]

References

[edit]

LXT: Interview With Cofounder & Chief Solutions Officer Mohammad Omar. Pulse 2.0. May 7, 2025.

LXT Acquires Clickworker in Data-for-AI Market Consolidation Move. Slator. December 17, 2024.

LXT Survey of Executives Reveals State of AI Maturity. TDWI. March 15, 2024.

LXT Releases The Path to AI Maturity 2023, An Executive Survey. AI Tech Park. March 15, 2023.

Use of AI for Training Data in Global Organizations. Crowdsourcing Week. [n.d.]

[edit]
This page will be placed in the following categories if it is moved to the article namespace.
Categories:
  1. ^ a b "LXT: Interview With Cofounder & Chief Solutions Officer Mohammad Omar". Pulse 2.0. April 2024.
  2. ^ a b "LXT Acquires Clickworker in Data-for-AI Market Consolidation Move". Slator. December 17, 2024.
  3. ^ "LXT Survey of Executives Reveals State of AI Maturity". TDWI. March 15, 2024.
  4. ^ a b "Use of AI for Training Data in Global Organizations". Crowdsourcing Week. Retrieved May 5, 2025.
  5. ^ "LXT Releases The Path to AI Maturity 2023, An Executive Survey". AI Tech Park. March 15, 2023.
Retrieved from "https://en.wikipedia.org/w/index.php?title=Draft:LXT_AI&oldid=1289418467"